Magnetic Resonance Imaging of Congenital Heart Disease (Moremedia) 3031292340, 9783031292347

Citation preview

Mushabbar A. Syed Raad H. Mohiaddin Editors

Magnetic Resonance Imaging of Congenital Heart Disease Second Edition

123

Magnetic Resonance Imaging of Congenital Heart Disease

Mushabbar A. Syed  •  Raad H. Mohiaddin Editors

Magnetic Resonance Imaging of Congenital Heart Disease Second Edition

Editors Mushabbar A. Syed Rolf & Merian Gunnar Professor of MedicineCardiology, Director Cardiovascular Imaging & Program Director, Cardiovascular Disease Fellowship Loyola University Medical Center Maywood, IL, USA

Raad H. Mohiaddin Professor of Cardiovascular Imaging Royal Brompton and Harefield Hospitals, Guy’s and St Thomas’ NHS Foundation Trust & National Heart and Lung Institute, Imperial College London London, UK

ISBN 978-3-031-29234-7    ISBN 978-3-031-29235-4 (eBook) https://doi.org/10.1007/978-3-031-29235-4 © The Editor(s) (if applicable) and The Author(s), under exclusive license to Springer Nature Switzerland AG 2012, 2023, corrected publication 2023 This work is subject to copyright. All rights are solely and exclusively licensed by the Publisher, whether the whole or part of the material is concerned, specifically the rights of translation, reprinting, reuse of illustrations, recitation, broadcasting, reproduction on microfilms or in any other physical way, and transmission or information storage and retrieval, electronic adaptation, computer software, or by similar or dissimilar methodology now known or hereafter developed. The use of general descriptive names, registered names, trademarks, service marks, etc. in this publication does not imply, even in the absence of a specific statement, that such names are exempt from the relevant protective laws and regulations and therefore free for general use. The publisher, the authors, and the editors are safe to assume that the advice and information in this book are believed to be true and accurate at the date of publication. Neither the publisher nor the authors or the editors give a warranty, expressed or implied, with respect to the material contained herein or for any errors or omissions that may have been made. The publisher remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. This Springer imprint is published by the registered company Springer Nature Switzerland AG The registered company address is: Gewerbestrasse 11, 6330 Cham, Switzerland

“To my family, Jennifer, Ameena, Aleena, Daneyal, and Cassie (and Newt— the best golden retriever) for their unwavering love and support and to my trainees (past, present, and future) for making work fun and productive.” —Mushabbar A. Syed “To the memory of my parents for their tremendous sacrifices, may they rest in eternal peace. To my wife, Khalida, and my children, Hasan, Zain, and Reema, for their love, support and understanding” —Raad H. Mohiaddin

Foreword

A CMR team experienced in and dedicated to congenital heart disease is integral to any Congenital Heart Disease Program As an Adult Congenital Heart Disease (ACHD) cardiologist who dedicated his career to ACHD, I read with great interest Magnetic Resonance Imaging in Congenital Heart Disease, and I feel honored and privileged to write a foreword to this landmark book. I congratulate the editors and the expert contributors to their second edition of their outstanding and comprehensive book. Cardiovascular Magnetic Resonance Imaging (CMR) provides highly valuable contributions to the diagnostic assessment, decision making, and management of congenital heart disease (CHD) patients. It is essential to accurately describe cardiac anatomy and to investigate (normal and abnormal) blood flow. CMR has become the reference imaging modality because of its highly reliable, reproducible data for the serial assessment which is crucial and more important than a single snapshot. CMR is the noninvasive imaging modality of choice to quantify ventricular volumes and ejection fraction, shunt calculations, and blood flow measurements in patient with complex CHD and multiple resources of pulmonary and/or systemic blood supply. It overcomes the concerning limitations of the Fick method to calculate the complex pulmonary blood supply in such patients and to estimate pulmonary vascular resistance. CMR has also become an integral diagnostic puzzle for risk stratification of sudden death. It is not appropriate to perform CMR in CHD patients as a “boutique business” in the current century anymore: CMR in CHD must be performed as a core business, with a team of imagers with expertise in CHD, who is embedded in a multidisciplinary team of a CHD program. This book sets these high-quality standards for how CMR is to be performed in CHD. The first chapter describes the general principles of cardiac CMR, including physics and CMR hardware, and makes this complex information understandable for non-radiologists; it also includes novel advanced techniques such as 4D flow analysis. This chapter extensively discusses the standard pediatric/congenital heart disease examination and protocols. It also highlights important tips and tricks to remain motionless in the CMR scanner and to perform a CMR under sedation or general anesthesia, a very challenging, but very important issue in the growing number of patients with syndromes and neurodevelopmental and neurocognitive deficits and disorders. The second chapter is dedicated to CMR Safety, including absolute and relative contradictions for patients with devices, CMR during pregnancy, stress medication for CMR, and many other issues related to this very safe diagnostic modality, but some potential risks. A special chapter then discusses the benefit of Gadolinium-based contrast agents to assess perfusion, characterize the tissue, and identify scar. Adverse reactions rarely occur despite the excellent safety profile. Chapter 4, Introduction to Congenital Heart Disease Anatomy, prepares the readers for the following chapters with detailed presentation of specific congenital heart defects (e.g., septal defects, tetralogy of Fallot, transposition of the great arteries, Ebstein anomaly). This key Chap. 4 describes the segmental approach in CHD that is fundamental and based on the founvii

viii

Foreword

dation of morphology: we always refer to morphology in CHD, we never refer to the position of a chamber (e.g., the right atrium and right ventricle can be positioned on the left, and they are still called right atrium and right ventricle, respectively). The editors and authors then present important CMR-related information about inherited cardiomyopathies, coronary artery anomalies, pericardial disease, cardiac tumors, and stress CMR in CHD. Fetal Cardiovascular Magnetic Resonance (Chap. 21) is another highlight of the book and addresses “investigation of fetal circulatory physiology in health and disease.” CMR is playing emerging and instrumental roles in the catheter laboratory (Interventional Cardiovascular Magnetic Resonance—Chap. 22) and in the electrophysiology suite (Magnetic Resonance in ACHD Electrophysiology—Chap. 23) to guide catheter-based interventions, to plan venous and arterial access, “to create 3D volume reconstruction and image integration with 3D electrophysiology mapping systems to facilitate retrograde approach and to avoid puncture of baffles in patients with atrial switch procedures or a Fontan circulation.” And last but not least, the chapter about 3D printing in CHD, including 3D CMR reconstruction for virtual tours through the heart, highlights the role of these novel, evolving, innovative techniques for teaching of trainees, interventionalists, and congenital heart surgeons for planning of catheter-based interventions and surgical repairs of complex congenital heart defects. In summary, the very informative book chapters are very well structured and organized. The text is supported by very informative tables, high-quality figures, and movies. The excellent schematic diagrams are very well designed and very illustrative for the readers. This book takes the readers on an exciting CMR journey of the broad spectrum of cardiovascular diseases with a special focus on CHD. Navigating through the book and the different chapters, I am reinforced in my strong opinion that CHD patients need to be imaged by radiologists with special expertise in CHD and/or CHD cardiologists with special training in CMR: cardiovascular imaging in CHD should not be offered as a “boutique business” anymore; it needs to be a core business in centers with expertise in complex CHD! The ideal organizational world of a CMR suite would combine the talents of the radiologists with CHD expertise and the talents of CHD cardiologists with CMR training to consolidate their experiences and expertise in dedicated teams! The readers highly appreciate the wonderful work of the editors and contributors who invested a tremendous amount of time to put together such an informative, very well-illustrated book. Dr. Erwin Oechslin, FESC FRCPC Cardiology / DRCPSC Adult Congenital Heart Disease Professor of Medicine, University of Toronto Peter Munk Cardiac Centre, Toronto ACHD Program Toronto, ON, USA

Foreword (First Edition)

As a cardiologist with a career interest in congenital heart disease in adults, I was delighted to have the opportunity to read Magnetic Resonance Imaging of Congenital Heart Disease and write a foreword for it. Over the past two decades, CMR has come to occupy an ever more important place in the assessment and management of patients with congenital heart defects (CHD) and other cardiovascular disorders. Thus, this new text will be of broad interest to both imagers and clinicians who deal with cardiovascular disease. Cardiovascular MRI offers an ever-expanding amount of information about the heart and circulation. It can provide outstanding images of cardiovascular morphology and function. It is increasingly being used to detect pathologic fibrosis and has an expanding role in the assessment of myocardial viability. Amazing though CMR is, its limitations and weaknesses also need to be clearly understood. As CMR has evolved, it has challenged other imaging modalities to improve and evolve, all of which improve the understanding that we clinicians have of our patients and consequently the care we can offer them. In many ways, cardiovascular CMR has offered important new insights that clinicians and imagers a generation ago would have considered impossible. The editors and their expert contributors have taken a large step toward making detailed information about CMR accessible to those working in the field and to those who use the information derived from CMR in their clinical practices. The field is ever changing and ever improving. This text offers an excellent foundation for the reader who is not familiar with the field and up to date descriptions of where we stand with imaging a broad range of cardiovascular diseases. The text is well referenced without being overwhelming, and the illustrations are generally outstanding, including many movie images. The opening chapter reviews the general principles of CMR, including information about the physics of the technique, and the hardware used. It reviews such useful subjects as remaining motionless in the scanner and the use of sedation and anesthesia. It provides important information relating to how to structure a study of various congenital cardiovascular conditions. The second chapter deals with the important issue of MRI safety, culminating (as in many chapters) with a series of “practical pearls.” Chapter 3 provides an introduction to the anatomy of CHD. The next seven chapters cover the subject of CMR in CHD in all its various forms. A particular emphasis is given to the role of CMR of septal defects, tetralogy of Fallot, Ebstein anomaly, transposition of the great arteries, and single ventricle/Fontan circulations. Chapter 12 focuses on aortic abnormalities including aortic coarctation, PDA, aortic aneurysms, and vascular rings. Chapter 13 deals with inherited cardiomyopathies, including hypertrophic cardiomyopathy, dilated cardiomyopathy, left ventricular non-compaction, and arrhythmogenic right ventricular dysplasia. Chapter 14 addresses coronary artery anomalies and discusses the appropriate roles of CMR and CT angiography in assessing such patients. The next two chapters deal with pericardial diseases and cardiac tumors. Chapter 17 discusses CMR in children, and Chap. 18 describes the current status of interventional CMR, an area with exciting potential. The final chapter explores the emerging roles for CMR in ACHD electrophysiology. These last two chapters represent an exciting marriage of differing imaging and therapeutic modalities that move the field forward.

ix

x

Foreword (First Edition)

Overall, the text offers the reader an exciting and comprehensive voyage through the place of CMR in a broad range of cardiovascular diseases with a special focus on congenital heart disease. It succeeds in describing the technical details of MRI techniques in sufficient detail to also help the clinician understand the most important elements of CMR in assessing and managing their patients. We readers are indebted to the editors and their contributors for having put together such an excellent and much needed text on this topic. Gary Webb

Preface

We are pleased to present the second edition of Magnetic Resonance Imaging of Congenital Heart Disease. We felt that the expanding role of MRI in congenital heart disease required a comprehensive reference text which led to the original book in 2013, and we were humbled and pleased with the reception it received. Since the first publication, there have been new advances in the field necessitating a revised and expanded new edition. Our publication schedule was significantly delayed due to the COVID pandemic; however, we took great care in assembling an expert panel of international authors to make this book as accurate and useful as possible. For the second edition, all previous chapters have been extensively reviewed and updated. In addition, we have added five new chapters on Gadolinium-Based Contrast Agents (Chap. 3), Pulmonary Hypertension (Chap. 8), Stress MRI in Congenital Heart Disease (Chap. 19), Fetal CMR Imaging (Chap. 21), and 3D Printing in Congenital Heart Disease (Chap. 24). We believe that this edition provides a contemporary and comprehensive review of the topic of MRI in congenital heart disease that will be useful to a wide variety of readers from trainees to expert professionals. Foreword to the original edition was written by Professor Gary Webb, a preeminent leader in the field of congenital heart disease who sadly died in 2021. We fondly remember him and his numerous contributions and accomplishments in this field that are an inspiration to us all. We are thankful to Professor Erwin Oechslin for writing the Foreword to the second edition. Professor Oechslin trained under Professor Webb in adult congenital heart disease at Toronto General Hospital/University of Toronto and later succeeded him as the Director of Toronto Adult Congenital Heart Disease Program. We are indebted to all of our chapter authors who believed in this project and ensured its completion despite their schedules impacted by the pandemic. We are also grateful to the Springer team led by Grant Weston for their patience and support throughout this project. Lastly, we are thankful to our readers, and hope that they enjoy reading this book as much as we enjoyed editing and authoring it. Maywood, IL, USA London, UK

Mushabbar A. Syed Raad H. Mohiaddin

This work contains media enhancements, which are displayed with a “play” icon. Material in the print book can be viewed on a mobile device by downloading the Springer Nature “More Media” app available in the major app stores. The media enhancements in the online version of the work can be accessed directly by authorized users. xi

Contents

1 General  Principles of Cardiac Magnetic Resonance Imaging �������������������������������   1 Mark Alan Fogel 2 MRI Safety�������������������������������������������������������������������������������������������������������������������  39 Roger Luechinger 3 Gadolinium-Based Contrast Agents�������������������������������������������������������������������������  51 Menhel Kinno and Joanne Sutter 4 Introduction  to Congenital Heart Disease Anatomy�����������������������������������������������  59 Pierangelo Renella and J. Paul Finn 5 Venoatrial Abnormalities�������������������������������������������������������������������������������������������  79 Henryk Kafka and Raad H. Mohiaddin 6 Septal Defects ������������������������������������������������������������������������������������������������������������� 103 Inga Voges and Sylvia Krupickova 7 Right Ventricular Anomalies������������������������������������������������������������������������������������� 115 Frédérique Bailliard 8 Pulmonary Hypertension������������������������������������������������������������������������������������������� 137 Daniel Knight and Vivek Muthurangu 9 Tetralogy of Fallot������������������������������������������������������������������������������������������������������� 147 Michael A. Quail, Vivek Muthurangu, and Andrew M. Taylor 10 Ebstein’s Anomaly and Other Tricuspid Valve Anomalies������������������������������������� 167 Steve W. Leung and Mushabbar A. Syed 11 Abnormalities  of Left Ventricular Inflow and Outflow������������������������������������������� 179 Tal Geva and Puja Banka 12 Single  Ventricle and Fontan Procedures������������������������������������������������������������������� 199 Sylvia Krupickova, Inga Voges, and Raad H. Mohiaddin 13 Transposition of Great Arteries��������������������������������������������������������������������������������� 213 Joel R. Wilson and Mushabbar A. Syed 14 Aortic Anomalies��������������������������������������������������������������������������������������������������������� 229 Sylvia S. M. Chen and Raad H. Mohiaddin 15 Inherited Cardiomyopathies ������������������������������������������������������������������������������������� 251 Theodore Murphy, Rory O’Hanlon, and Raad H. Mohiaddin 16 Coronary Artery Anomalies��������������������������������������������������������������������������������������� 273 Andrew Crean 17 Pericardial Diseases ��������������������������������������������������������������������������������������������������� 303 Edward T. Martin xiii

xiv

18 Cardiac Tumors����������������������������������������������������������������������������������������������������������� 315 Mushabbar A. Syed and Raad H. Mohiaddin 19 Stress  MRI in Congenital Heart Disease������������������������������������������������������������������ 331 W. A. Helbing 20 Paediatric CMR ��������������������������������������������������������������������������������������������������������� 347 Emanuela R. Valsangiacomo Buechel 21 Fetal  Cardiovascular Magnetic Resonance ������������������������������������������������������������� 361 Adrienn Szabo, Liqun Sun, and Mike Seed 22 Interventional  Cardiovascular Magnetic Resonance����������������������������������������������� 383 Vivek Muthurangu, Oliver Richard Tann, and Andrew M. Taylor 23 Emerging  Roles for Cardiovascular Magnetic Resonance in Adult Congenital Heart Disease Electrophysiology������������������������������������������������� 397 Sophie A. Jenkins, Jennifer Keegan, Sabine Ernst, and Sonya V. Babu-Narayan 24 3D  Printing in Congenital Heart Disease����������������������������������������������������������������� 415 Michael D. Seckeler, Claudia E. Guerrero, and Andrew W. Hoyer  Correction to: 3D Printing in Congenital Heart Disease��������������������������������������������������C1 Michael D. Seckeler, Claudia E. Guerrero, and Andrew W. Hoyer Index������������������������������������������������������������������������������������������������������������������������������������� 429

Contents

1

General Principles of Cardiac Magnetic Resonance Imaging Mark Alan Fogel

1.1 Introduction A basic understanding of the underlying principles of cardiovascular magnetic resonance imaging (CMR) and methods used to form images is important if one is to successfully image in clinical practice or research and interpret the data correctly. This section will provide a brief overview of the fundamentals and some techniques in CMR imaging. For more information, the reader is referred to the references in this chapter or the larger textbooks on fundamentals of magnetic resonance imaging as well as other chapters in this book [1].

1.2 Physics and CMR Hardware The crux of CMR is nuclear magnetic resonance where a signal is emitted by a sample of tissue after radiofrequency energy is applied to it. Note that this signal is emitted by tissue molecules in contrast to X-ray imaging where the tissue or contrast agents attenuate externally applied radiation. At the atomic level, it has been well known that spins and charge distributions of protons and neutrons generate magnetic fields. Only certain nuclei can selectively absorb and subsequently release energy since it requires an odd number of protons or neutrons to exhibit a magnetic moment associated with its net spin. The hydrogen atom is the one utilized in CMR imaging since it consists of a single proton with no neutrons (which gives it a net spin of ½), its large magnetic M. A. Fogel (*) Pediatrics (Cardiology) and Radiology, Perelman School of Medicine at The University of Pennsylvania, Philadelphia, PA, USA Director of Cardiac Magnetic Resonance, The Children’s Hospital of Philadelphia, Philadelphia, PA, USA e-mail: [email protected]

moment and its abundance in the body (water and fat). Although each magnetic moment of individual hydrogen protons themselves is small, because of the abundance of that atom in the body, the additive effect of the many magnetic moment vectors makes it detectable in CMR. Generally, the net magnetization of a tissue in the body is zero as there is a random orientation of the individual protons or “spins”; stochastically, the odds greatly favor a zero magnetization. However, when the body is placed in a strong magnetic field (Fig. 1.1) such as 1.5 or 3 T MRI systems (for comparison, the Earth’s magnetic field is approximately 0.05 mT at the surface), the spins align themselves with the applied field either parallel or anti-parallel to the field. In addition, the atoms undergo a phenomenon known as precession (such as the motion of a spinning top as it loses its speed) whose axis is based around the direction of the magnetic field (Fig. 1.1a); this precession, described as cycles per second, is described by the most famous equation in CMR and MRI—the Larmor equation, ω = γB0, where ω is the frequency of precession of protons in an external magnetic field, γ is a constant called gyromagnetic ratio, and B0 is the external magnetic field (the magnetic field generated by the MRI system). There is a different gyromagnetic ratio for each atom; for hydrogen, it is 42.58 MHz/T which generates a frequency of approximately 64 MHz at 1.5 T (Larmor frequency for a 1.5  T magnet is 1.5 (T)  ×  42.56 (MHz/T)  =  63.8  MHz). When a radiofrequency pulse is applied which just happens to match the Larmor precessional frequency, some of the protons will flip to a high energy state. For protons at field strengths used for CMR, radiofrequencies in range of “very high frequency” or VHF can be used which is non-ionizing, contributing to the inherent safety of MRI when compared to X-rays. To get from here to how a signal is generated from tissue, two more concepts must be introduced. As mentioned, the hydrogen spins are either in the low or high energy spin states with only slightly more spins in the low energy state. The number of excess spins is directly proportional to the

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. A. Syed, R. H. Mohiaddin (eds.), Magnetic Resonance Imaging of Congenital Heart Disease, https://doi.org/10.1007/978-3-031-29235-4_1

1

2

a

M. A. Fogel B0

B0

z

M

y

x

b

B0

c

z

1

z

2

+ve gradient

x

x

M

y

y B1

3

z

4

z

–ve gradient

x

x y

y

Fig. 1.1 (a) Protons spin and process like a top wobbling (left). If the proton is at the (0, 0, 0) coordinate of an x, y, z coordinate system (right), its axis is represented by the blue vector M and wobbles around the z axis which is in line with B0 at a frequency ω. (b) After energy is inputted into the system, the axis flips (in this particular instance, 90°) and then slowly returns to its original position. (c) In (1) the protons are

flipped 90 with subsequent “dephasing” of the spins in (2) which can be increased by a gradient (i.e., faster protons and slower protons will separate). In (3), energy can be inputted into the system to flip the protons to an exact mirror image so that the faster spins are not behind the slower spins. Finally, the faster spinning protons catch up to the slower spins to create a large detectable signal (in (4))

total number of spins in the sample and the energy difference between states (Boltzmann equilibrium probability). The formula used to determine this difference is N−/N+  =  e−E/kT where N− is the number of spins in high energy state, N+ is number of spins in the lower energy state, k is Boltzmann’s constant (1.3805  ×  10−23  J/K), T is the temperature (in Kelvin), and E is the energy difference between the spin states. The second concept is that the energy of a proton (E) is directly proportional to its Larmor frequency υ (in Hz), such that E  =  hυ, where h is Plank’s constant (h = 6.626 × 10−34 J s). By substituting into the Larmor equation, this yields the relationship between E and the magnetic field B0, E = hγB0. When energy is inputted into the system

and it matches the energy difference between the lower and higher energy spin states, atoms from the lower energy states get flipped up to the higher energy states. As these atoms then return to the lower energy state, they release energy and this signal, the resonance phenomenon, can be detected (Fig. 1.1b). This is how the MRI signal is generated. It follows that only those excess spins in the low energy state can be excited to the high energy state and generate the MRI signal. It is amazing that there are only approximately nine more spins in the low energy state compared to the high energy state for each two million spins at 1.5 T! However, one must also realize that since each mL of water contains nearly 1023 hydrogen atoms, the Boltzmann distribution

1  General Principles of Cardiac Magnetic Resonance Imaging

1.2.1  T1 Relaxation When the protons are flipped to transverse plane, the Mz component of magnetization decreases to near zero (Fig.  1.1b, c); the time for return of this magnetization Mz after the RF pulse is turned off is measured by the time constant T1 which is defined as the time necessary to recover 63% of the equilibrium magnetization M0 after the 90° RF pulse (Fig. 1.2): Mz(t) = M0(1 − exp(−t/T1)). Physically, the return of longitudinal magnetization is a function of how fast the spins release their energy to the tissue which is termed the “lattice” hence T1 being called spin-lattice relaxation. As one might guess, this process depends in part on the physical properties of the tissue where the frequency of precession of the spins needs to overlap the frequencies of the molecules in the lattice. In addition, the process is also dependent on the

T2 Decay Transverse Mo

T1 Recovery

63% Aligned Mo

d­ iscussed above predicts over 1017 spins contributing to the MRI signal in each mL of water! It is interesting to note that the higher the magnetic field strength, the greater the number of excess spins in low versus high energy state; it follows that as the field strength increases, so does the magnitude of the MRI signal. Hence, there is a push by manufacturers to create larger and larger magnetic fields from the 1.5 T fields most commonly used today. Indeed, many 3 T systems have been deployed and 7 T systems have been discussed for clinical use. It should also be noted, as well, that there is research in the opposite direction into lower field systems such as 0.55 T. With modern technology such as advanced gradient systems, pulse sequences, computers, and software, high-­ quality CMR can be performed at that field strength close to the quality of higher field strength systems [2, 3]. When the radiofrequency energy is applied that matches the Larmor frequency, some of the protons in the low energy state jump up to the high energy level as noted above. This radiofrequency pulse (RF) has a magnetic field itself, B1, which is perpendicular to the direction of B0. It is of mT order of magnitude and tilts the longitudinal magnetization (Mz) a certain amount depending upon the duration of the RF pulse and the strength of B1 field. If an RF pulse is applied to tilt the net magnetization from the longitudinal plane (Z plane) totally to the transverse (XY) plane (called a 90° RF), the transverse component of the net magnetization is the one that will generate an induced voltage in a receiver antenna (the MR signal). The way this occurs is through what is termed “relaxation” where the protons return from their excited state to a low energy state (Fig. 1.1b). The duration of the induced voltage is a function of the time it takes to undergo relaxation and is described by relaxation time constants termed T1 and T2 which describe the changes in longitudinal magnetization (Mz) and transverse magnetization (Mxy) respectively.

3

37% time

T1

time

T2

Fig. 1.2  T1 relaxation or recovery (left) is the return of magnetization Mz, to the equilibrium magnetization M0; the time constant is defined as recovering to 63%. T2 relaxation or decay (right) is the decay in the transverse plane

main magnetic field strength; at higher fields, the frequencies of spins precession increase with less overlap of frequencies in the lattice, resulting in a longer T1. Water, however, has a frequency range that is large.

1.2.2  T2 Relaxation When the protons are flipped to transverse plane, the Mxy component of magnetization becomes maximized as all the protons precess with the same phase, called phase coherence (Fig. 1.1b, c). The spins of each proton in the general vicinity of each other interact with each other, however, and in time, this coherence is lost resulting in a decrease in the net magnetization (Fig.  1.1c) and induced voltage in the receiver antenna. This is appropriately called spin-spin relaxation or transverse relaxation and is measured by the T2 time constant. The transverse magnetization (Mxy) will decay exponentially from Mz by the following formula (Fig.  1.2): Mxy(t)  =  Mz(0)exp(−t/T2). The time constant is defined as decaying to 37% of its initial value. This relaxation is highly dependent on the makeup of the tissue; small molecules in a generally unstructured tissue have long T2 values because fast and rapidly moving spins average out the intrinsic magnetic field inhomogeneities while large molecules in densely packed tissue have shorter T2 values. Unfortunately, there are other factors responsible for decay of magnetization in the transverse plane; imperfections of the main magnetic field, susceptibility differences between nearby tissues can and do contribute to the loss of phase coherence (Fig. 1.1c). This is measured by the time constant T2*. In general, T1 is always greater than T2 which is always greater than T2*.

1.2.3 Image Formation Now that the basic physical properties are defined, the discussion can turn to creating images. To create images, a

4

M. A. Fogel

­ agnetic field gradient must be formed. Within the main m magnetic field, B0, all protons precess at the same frequency (Fig. 1.1a). The Larmor equation tells us that this precession is a function of this field strength; by changing the magnetic field ever so slightly by position and time, the precession of the protons can be changed ever so slightly by position and time. Using this information, localization of the MR signal from the precise part of the body can be accomplished and images can be generated. This precision controlled alteration of the magnetic field is created by gradient coils, which generate linear variations in the main magnetic field strength in three orthogonal planes (Figs. 1.1c and 1.3a). By using these coils simultaneously, a linear magnetic field gradient can be generated in any direction. This gradient changes the precession frequency of the protons at precise locations in a linear fashion. To select a certain plane (slice) in the body (Fig. 1.3b), an RF pulse is applied and it follows that if the RF pulse center frequency is shifted to match a specific location along the gradient, it will selectively excite the protons at that region. A slice of arbitrary thickness, orientation, and location along the direction of the “slice select gradient” can therefore be selectively excited to generate the signal used to form the MR image and the signal detected by the MRI receiver coil will come from the excited slice only

a Gradients in the magnetic field

b

y x z

Fig. 1.3  By altering the magnetic field in three orthogonal planes, gradients can isolate a plane in space (a). This is accomplished by flipping the protons only in the plane desired (b)

(Fig. 1.3b). The amplitude of the signal is directly proportional to its thickness, practically limiting the thickness at approximately 2 mm. After selecting the slice, the image itself needs to be created in two dimensions in the xy plane (practically speaking in 2 orthogonal planes of the bore of the magnet—right/left and up/down when looking into the bore). This creates the pixels (2-dimensional picture elements); in 3-dimensional imaging, this is called voxels (3-dimensional volume elements). As with choosing the slice, linear field gradients and the Larmor relationship between field strength and precessional frequency are used to encode spatial location information into the MRI signal. After a slice-selective RF pulse, a linear magnetic field gradient is switched on in one of the in-plane directions of the image, perpendicular to the “slice select gradient”; this gradient changes the precessional frequency in a linear distribution along the gradient direction allowing the identification of every location along the gradient by the frequency of the signal (Fig. 1.3). This is called frequency encoding. The MR signal is detected and put through an analog-to-digital converter; remembering that we have encoded the slice and one direction in the plane of the image at this point, the signal is thus the amalgamation of all of these frequencies. Therefore, the signal varies with position, also called its “spatial frequency”; this is called “k-space.” If one looks at the distribution of the signal, it creates a sinusoidal distribution of phase across the direction of the gradient; this describes a single spatial frequency kx (important in phase encoding in the next step). A special mathematical technique called the “Fourier transformation” is used to separate out the individual frequency components in the detected signal, decoding the signal into individual signals coming from locations along the frequency encoding gradient. The Fourier transformation can be used to translate the signal from “k-space” to the image and vice versa (Fig. 1.3). Finally, the third spatial dimension (second in-plane dimension) must also be encoded (if frequency encoding is the “x,” the “y” in the “xy” plane must also be created); the technique used is called “phase encoding” (Fig. 1.4) and is also based on the Larmor equation. Phase encoding is accomplished with the application of a number of gradient pulses of differing amplitudes which encode a specific spatial frequency, ky (Fig.  1.5). This phase encoded gradient pulse is changed to encode a different spatial frequency component prior to each frequency encoding gradient. In this manner, by successive phase encoding pulses, and “y” part of the image is built up and a matrix is formed; this matrix is referred to as k-space, and the numerous gradient pulses “fill” the k-space until the image is complete. The 2-dimensional Fourier transform is utilized to convert the spatial frequencies created by the phase encoding steps into an image (Fig.  1.3). Phase encoding can also be used in the slice direction to

1  General Principles of Cardiac Magnetic Resonance Imaging

5

Phase Encoding

180°

α°

RF SLICE

PHASE

Readout gradient field strength

Magnitude

Samples

+

=

Detected RF Signal

Emitted RF Signals

READOUT

Fourier Frequency Transform

Fig. 1.4  The image is created by a combination of frequency encoding in one direction (in this case, in the x plane or the horizontal portion of the image) and phase encoding in the other dimension (in this case, in the y plane or the vertical portion of the image). For frequency encoding, a gradient is applied to change the magnetic field in that direction (right triangle). Different frequencies correspond to different positions (different colors and waves on the diagram) which produce a detected radiofrequency (RF) signal which is a combination of all the frequencies of the various positions (rightmost signal). When put through the Fourier transform, signals can be separated into their different positions (lower graph). The vertical portion of the image is created by changing phases of the radiofrequency pulses (phase encoding, see text). The cardiac magnetic resonance image is of a “4-chamber” view of a patient with single ventricle after Fontan

encode thinner slices than possible with the slice selection gradient technique, so-called 3D data acquisition.

1.2.4 MRI Hardware Once all the components of generating an MR signal are known, it is important to review the equipment needed. There is of course the main magnet (B0 or Bz field), the RF transmitter coil (B1 field), the gradient coils (Gx, Gy, Gz fields), and receiver coils which “listen” for the signal. In addition, there are second-order shim coils which are often used to achieve a more homogeneous B0 field. There are a number of computer systems including those which are used to control the MRI magnetic field-generating units and those used to reconstruct the acquired data. There is also a system which provides an interface for the user and the other systems.

ADC SIGNAL

Fig. 1.5  A typical pulse sequence diagram includes (a) when the radiofrequency (RF) pulse is applied (top line) and at times, how much of a flip angle (in this case, 180°), (b) when the slice selection gradient (Gz or slice) is turned on (second line), (c) when the phase encoding gradient (Gy) is turned on (third line), when the frequency encoding gradient (readout or Gx) is turned on (fourth line), and when the analog to digital converter is turned on (fifth line) and the signal is created (sixth line). This is an example of a magnetization prepared gradient echo pulse sequence

1.2.5 Pulse Sequences This is the sequence of events which control all the various factors involved in the creation of an image. It is important to note that these times are on a microsecond scale and need to be controlled by computer for precise timing. Timing of all the gradients switching on and off, the phase encoding, the RF pulses, the analog to digital converting data sampling, and control of transmitter and receiver operation are all defined by the pulse sequence. As there are a limitless amount of pulse sequences, it is impossible to describe all of them; however, to understand them, a pulse sequence diagram is used which details the timing of each component; a representative pulse sequence diagram is shown in Fig. 1.5. To simplify the concepts, it should be noted that there are five broad concepts with regard to pulse sequences which may be understood to aid in examining many of the pulse sequences in use. They are as follows: 1. Magnetization preparation is a technique employed, usually at the beginning of the sequence, which changes the tissue characteristics prior to actually creating the image (Fig.  1.5). T2 preparation, for example, can be employed to suppress myocardial muscle and is used in visualizing coronary arteries (Fig.  1.6). The inversion recovery technique uses a 180° RF pulse to magnify differences in different tissue characteristics of T1; the saturation recovery technique uses a 90° RF pulse prior to image. The inversion recovery technique is used in

M. A. Fogel

6 Fig. 1.6 (a) T2 prepared steady-state free precession to visualize coronary arteries. The image on the left demonstrates a right coronary artery (RCA) giving rise to a left circumflex coronary artery (LCx). The image on the right demonstrates a coronary artery aneurysm from Kawasaki’s disease from the RCA. (b) ECG-gated contrast-enhanced inversion recovery gradient echo imaging of the right (upper left) and left (upper right) coronary arteries with a 3D reconstruction (lower image)

a

RCA

LCx

aneurysm

RCA

b RCA

RCA

LMCA and bifurcation

LMCA and bifurcation

7

1  General Principles of Cardiac Magnetic Resonance Imaging Fig. 1.7 Delayed enhancement imaging from a “4-chamber” (left) and long axis view (right) of a patient with hypoplastic left heart syndrome after Fontan. White arrows in red outline demonstrate some of the areas of myocardial scarring

RV

Baffle

LV

delayed-enhancement (Fig. 1.7), T1 mapping as well as 3. Filling k-Space: As noted above, k-space is filled with dark blood imaging [4] (Fig. 1.8) while saturation recoveach phase encoding step. Most sequences employ what ery is used in T1 mapping as well as first-pass perfusion is known as Cartesian k-space sampling, where there is a imaging (Fig. 1.9). linear filling of k-space with each phase encoding step. 2. Echo Formation: This is the “echo” referred to above and There are, however, other methodologies which have various types of echo formation is used in CMR.  An come into existence and are used. A “radial” filling of older technique of echo formation which still has applik-space has some advantages over Cartesian sampling cability today is called spin-echo, which is used most when it comes to efficiently filling the matrix and has often in dark blood imaging for morphology and tissue been used in cine imaging [11]. A “spiral” filling of characterization (e.g., myocardial edema) (Fig.  1.8). k-space trajectory has been utilized for coronary artery Another technique of echo formation, gradient-echo imaging, because it has some advantages in speed and imaging (Fig. 1.10), is used in a whole host of applicamost importantly, insensitivity to motion; unfortunately, tions such as: it is highly sensitive to field inhomogeneities and, there (a) Cine imaging for cardiac function including myocarfore, has not gained in popularity. A “radial” filling of dial tagging (Figs. 1.11 and 1.12) k-space can be used as well, collecting k-space in a series (b) Assessing valve morphology (Fig. 1.13), valve regurof spokes, with the “golden angle” technique being the gitation as well as valve or vessel stenosis most efficient and robust way to do so [12]. (c) In delayed-enhancement (Figs.  1.7 and 1.10) and 4. Segmentation: This refers to the number of lines of first-pass perfusion (Fig. 1.9) for myocardial scarring k-space filled per cardiac cycle [13]. If one line of k-space and myocardial perfusion respectively is filled per cardiac cycle, that pulse sequence is said to be (d) Phase contrast velocity mapping to determine blood “segmented”; if all the lines are filled in one cardiac flow (Fig. 1.14) cycle, that pulse sequence is said to be “single-shot” or Gradient echo imaging comes in couple of forms: (a) “non-segmented.” There are of course gradations of segunbalanced gradient echo imaging and (b) balanced mentation between the two and the degree of segmenta­gradient echo imaging also known as steady-state free tion is referred to by the number of lines of k-space filled precession imaging (SSFP) (Figs. 1.15, 1.16, and 1.17). per heartbeat (e.g., 3, 5, 7 segments or views, etc.). Any SSFP is more commonly used than the unbalanced gradilevel of segmentation can be used with any of the methent echo form due to its high signal to noise, blood to ods of magnetization preparation, echo formation, or the myocardium contrast, and imaging efficiency [5]. Echo-­ ways of filling k-space. It generally follows that if the planar imaging is used as a method for perfusion imaging more lines of k-space filled in a heartbeat, the less time it due to its high efficiency. will take to form the image while the reverse is true with

M. A. Fogel

8 Fig. 1.8 (a) Multiparametric mapping. Upper right and left images are T1 maps before and after gadolinium administration respectively where each pixel’s intensity represents T1 relaxation time. Using the hematocrit, an extracellular volume map can be created (lower left) where each pixel’s signal intensity represents the extracellular volume percentage. The lower right panel is a T2 map where each pixel’s intensity represents T2 relaxation time. (b) Dark blood imaging in a patient with a left ventricular hemangioma (tumor). The upper left panel is a dark blood T1-weighted image of the tumor demonstrating slight hyperenhancement; note the fat in the chest wall. The upper right panel is the same dark blood T1-weighted image with a “fat saturation” pulse applied prior to imaging (prepulse). Note how the fat in the chest wall is not present in this image because the prepulse destroyed all the spins of the fat; however, the tumor is still present, indicating there are no fatty elements in the tumor (it is not a lipoma). The lower left panel demonstrates hyperenhancement on T1-weighted imaging after gadolinium administration. The lower right image is a T2-weighted image with fat saturation demonstrating hyperenhancement indicating increased water content

a

RV

tumor

fat

tumor

LV

LV

b

Native T1 Map

Contrast Enhanced T1 Map

T2 Map

Extracellular Volume Map

1  General Principles of Cardiac Magnetic Resonance Imaging Fig. 1.9  Perfusion imaging in the patient in Fig. 1.8 with a left ventricular hemangioma (tumor). Four frames during the first pass myocardial perfusion imaging demonstrate contrast in the right ventricular (RV) cavity (I), the left ventricular (LV) cavity (II) where the tumor can be seen, and phases III and IV where the contrast enters the LV myocardium. Note how the tumor becomes signal intense in III and IV; so much so that it is indistinguishable from the cavity (somewhat in III and indistinguishable in IV). The tumor can be visualized in II as the contrast is in the cavity but not in the myocardium where presumably the tumor receives its blood supply from

Fig. 1.10 Tissue characterization in the patient in Figs. 1.8 and 1.9 with a left ventricular hemangioma. The upper left panel is a steady-­ state free precession image in short axis clearly showing the spherical tumor while the upper right panel is an axial gradient echo sequence also clearly showing the tumor. Note the tumor is isointense with cardiac muscle on steady-state free precession and the hypoenhancement on gradient echo imaging. The lower image is a delayed enhancement image of the tumor demonstrating hyperenhancement

9

I

tumor

II

RV

LV

III

IV

tumor

LV

tumor LV

M. A. Fogel

10 Fig. 1.11 Myocardial tagging in the “3-chamber” (upper and lower left), short axis (upper right), and long axis views (lower left) of a patient with a single left ventricle after Fontan. This is an example of spatial modulation of magnetization (SPAMM) where a grid is laid down on the myocardium and deformation can be visualized [6–9]. Note the deformation from end-­diastole (ED, upper left) to mid-systole (lower left) in the “3-chamber” view. It is the equivalent of speckle tracking in echocardiography except by CMR; the “speckles” are purposefully created in a certain geometry for strain and wall motion assessment

ED

LV

“3-chamber”

Baffle

short axis

LV

long axis

Mid-systole

Fig. 1.12 Myocardial tagging in the “4-chamber” (left) and long axis of the right ventricle (right) in a patient with pulmonic stenosis after balloon dilation. Note how this differs from Fig. 1.11 with one set of parallel lines laid down on the myocardium—the so-called 1-dimensional tagging [10]

RV long axis

4-chamber

LV RV

11

1  General Principles of Cardiac Magnetic Resonance Imaging Fig. 1.13  Gradient echo imaging (left) and through plane phase-encoded velocity mapping (right) of the aortic valve. Note how the right (RCC), left (LCC), and non-coronary commissures are easily visualized on both images, demonstrating the trileaflet valve

RCC

LCC

NCC

the less lines of k-space per heartbeat (i.e., the number of segments inversely proportional to the time it takes to create the image). It is also true, however, that the more lines of k-space acquired per heartbeat, the worse the temporal resolution will be (i.e., the number of segments inversely proportional to the number of time points that can be created in the cardiac cycle). Tradeoffs are part of CMR and for most applications such as phase contrast velocity mapping and flow, image creation can be obtained in a breath-hold. It is important to realize that a regular cardiac rhythm is needed to ensure that lines of k-space from each cardiac cycle is acquired during the same point in time of the cardiac and respiratory cycles. In patients with arrhythmias or an inability to breath-hold, single-shot or real-time methods are commonly used [14] although there are also arrhythmia rejection techniques whereby an R-R interval and target heart rate are set and only beats falling within that range are accepted to build the image. For those patients who cannot hold their breath, using segmented techniques with multiple “averages” or “exci-

tations” can be used to smooth out the respiratory motion at a minor cost to image fidelity (most times). 5. Image Reconstruction: As mentioned above, the Fourier transformation is used to create an image from the lines of k-space which is acquired from the MR signal. A technique called “partial Fourier” or “partial k-space” has been used for many years which reduces scan time with a lower signal-to-noise than using a “full” Fourier transformation. Parallel imaging with names such as SENSE [15], SMASH, GRAPPA, and TSENSE [16] uses multiple coils and multiple channels and has become ubiquitous in many sequences; they sample only a fraction of the full k-space but yet allow for a full field-of-view and resolution images with significant time savings at the cost of signal-to-noise. Compressed sensing CMR is another method that undersamples k-space losing very little fidelity to speed up the image acquisition and to reconstruct images that are nearly indistinguishable from those that do not undersample k-space [17].

M. A. Fogel

12 Fig. 1.14 (a) In-plane phase-encoded velocity mapping in a patient with double outlet right ventricle after a right ventricle to pulmonary artery conduit. On the left is a magnitude image of the right ventricular (RV) outflow tract (RVOT). In the middle is a mid-systole frame of the in-plane phase-encoded velocity map in the exact orientation and position as the image on the left where flow cephalad is signal intense; on the right is a mid-systolic frame where flow caudad is signal poor (dark on the image) indicating severe conduit insufficiency. (b) 4D flow in systole (right) and diastole (left) of a single ventricle patient after Fontan. Note the fenestration flow (arrows) and the vortex formation in the dilated ascending aorta (AAo)

a

RVOT

RV

b

Fontan

mid-systole AAo

Fontan

mid-diastole AAo

1  General Principles of Cardiac Magnetic Resonance Imaging Fig. 1.15  Steady-state free precession imaging in the 4-chamber view (left) at end-diastole (ED, upper) and end-systole (ES, lower) in a patient with tetralogy of Fallot. The long axis of the right ventricle (RV) is on the right. LV left ventricle

13 ED LV

RV

RV long axis

ES

Fig. 1.16  Steady-state free precession imaging in the short axis view at end-diastole (ED, left) and end-systole (ES, right) in the patient with tetralogy of Fallot in Fig. 1.15

End-diastole

RV

LV

End-systole

14 Fig. 1.17  Various steady-­ state free precession cines of the patient in Figs. 1.15 and 1.16 with tetralogy of Fallot. The right ventricular outflow tract views are seen in the upper panels in off-axis sagittal (left) and coronal (right) views which are orthogonal to each other. The right (RPA) and left pulmonary arteries (LPA) are seen in long axis in the left lower and right lower panels respectively. MPA main pulmonary artery, RV right ventricle

M. A. Fogel

MPA MPA

RV

RPA

MPA

LPA

RV

1.3 Prospective Triggering/Retrospective Gating Because the heart needs to be at the same phase of the cardiac cycle with any segmented technique, as noted above, a way is needed to determine this phase. This is nearly universally the R-wave of the ECG. A static or non-moving image uses the R wave to signal the beginning of systole as is the touchstone of the cycle; the CMR sequence then begins. This technique is called prospective triggering since the sequence is initiated by the R wave; lines of k-space are then acquired. Phases of the cardiac cycle are defined by a fixed time after the R-wave, so small perturbations of rhythm will put the heart at a slightly different point in the cardiac cycle; this generally does not affect the image too much. In addition, there is generally some “dead space” prior to the next R wave so very late diastole is usually not imaged or utilized. Cine or moving images are acquired by either this method or the method of retrospective gating. With retrospective gating, lines of k-space are acquired continuously regardless of the phase of the cardiac cycle while the ECG is simultaneously recorded; after image acquisition, the software “bins” the lines of k-space relative to the ECG and cardiac cycle. In this way, each cardiac phase is defined as a certain percentage of the cardiac cycle, allowing the actual duration of each phase to vary flexibly with variation in cardiac cycle. In addition,

no “dead space” is left prior to the next phase which can be important in assessing flows or ventricular function. The above paragraph makes a distinction between static and dynamic techniques. Static ones are generally used for cardiovascular anatomy or characterizing tissue. Dynamic techniques are used to assess function or flow in addition to anatomy. A run of single-shot images, acquired quickly, can be strung together as motion and this is termed “real-time” and is asynchronous with the cardiac cycle; this can be used in cine imaging, phase contrast velocity mapping, or dynamic 3D angiography. First-pass perfusion imaging can be thought of as a hybrid between static and dynamic imaging, where each image depicts a different phase of the cardiac cycle over time.

1.4 ECG Signal For many years, the upstroke of the R wave on the ECG signal was used to trigger the scanner and used as a marker for end-diastole; unfortunately, artifacts occurred because of the high magnetic field strength and radiofrequency pulses which precluded reliable detection of the true R wave. Bizarre T waves and spikes during the ST segment of the ECG would cause the triggering to falsely detect these waves as the R wave. This is especially true in congenital heart dis-

1  General Principles of Cardiac Magnetic Resonance Imaging

ease where abnormal QRS axes and bundle branch blocks from surgery can make the distinction of the R wave even more problematic in the scanner. On most systems in use today, detection of the R wave as a trigger has been replaced by the use of vectorcardiography (VCG), which is less susceptible to distortion from the magnetic field and of flowing blood in the thoracic aorta which can act as a conductor. Although wired connection between the VCG and the imaging systems has been utilized in the past, this has been nearly universally replaced by wireless transmission which allows for more flexibility in the scanner. There are alternatives to the direct connection between the “MRI ECG” and the patient. The ECG signal from external monitoring systems such as the anesthesia equipment can be used which can generate a signal contemporaneous with the R-wave to the MR scanner. Alternatively, the ECG signal can be discarded for “peripheral pulse triggering” where a finger or ear pulse may also be used; obviously, this requires good peripheral circulation. If the patient is cold or has a coarctation, this will often be unsuccessful. It also should be noted that because there is a delay in transmission of the pulse to the distal part of the body that is being monitored, the waveform will be delayed by 200– 300  ms when compared with the ECG; this needs to be taken into account during the analysis and interpretation phase of the examination. Peripheral pulse gating is especially useful in patients where a good detection of the R wave cannot be obtained otherwise. Another alternative is the use of non-triggered SSFP sequences where lines of k-space are continually being obtained by the imaging system without regard to the ECG or respiration (see above and below). With the use of parallel imaging in the spatial or time domains, a temporal resolution as high as 30–40  ms can be acquired; even higher temporal resolution can be obtained by combining this with compressed sensing. This type of imaging can be used in patients with arrhythmias to obtain functional information when triggering the ECG is problematic (see below). Finally, recent advances in hardware and software have enabled the use of “self-gating” sequences, where a coil is used to monitor the motion of the ventricle which is used as a signal for ventricular contraction and relaxation. This approach allows the heart itself to be monitored and act as its own signal for the imager; retrospective analysis of the lines of k-space can then be “binned” to construct moving images. There are now techniques that not only “self-gate” but also compensate for respiration where a 3D cine data set can be acquired at multiple respiratory phases without an ECG or navigator (see below) [18]. A special note is required on patients with arrhythmias. With frequent premature ventricular contractions, runs of supraventricular tachycardia or trigeminy for example, it is unclear what an ejection fraction, cardiac index, or end-­ diastolic volume would mean given that these ventricular

15

performance parameters can change from beat-to-beat. A qualitative assessment using real-time steady-state free ­precession is one way to get a handle on ventricular function. Nevertheless, there may be instances when some quantitative information may be needed; in these particular cases, “arrhythmia rejection” can be used (see above). With this approach, a range of heart rates or R-R intervals can be set, and the imaging system will only allow those lines of k-space which meet these requirements into the final image; the rest of the lines of k-space which fall outside these heart rates are ignored. This approach is inefficient, however, in this manner, quantitative ventricular performance information can be obtained for a range of heart rates. For example, if the range is set between an RR of 700 and 800 ms, the resulting cardiac index can be said to be present for heart rates between 75 and 86  beats/min. Finally, some real-time cine sequences have arrhythmia compensation built in [14].

1.5 Respiration Besides cardiac phases, respiration must be dealt with as it causes positional variation of the heart from movement of the lungs and diaphragm; if not taken into account, this will lead to motion artifacts. There are a number of ways in which this is dealt with in CMR: 1. Breath-holding, where the patient’s breath is held during image acquisition. For many common applications such as cine and phase contrast velocity mapping, image acquisitions are fast enough to be performed in a reasonably short breath-hold. This can be done in adults or in children under anesthesia who are paralyzed, intubated, and mechanically ventilated. These pulse sequences are widely available and commonly used routinely. 2. Signal averaging, also termed multiple excitations, where the signal from the complete image is “averaged” over many respiratory and cardiac cycles, “averaging out” the respiratory motion and making the image sharper than without this technique but less sharp than breath-holding. This can be used in small children unable to voluntarily breath-hold or adults who cannot cooperate. It has the advantage of being more “physiologic” and representative of the true state of the patient’s physiology as the patient is continually breathing the information is averaged over many respiratory cycles. 3. Respiratory gating, where the motion of the diaphragm or the chest wall is tracked by either a navigator pulse (which tracks diaphragmatic motion, the equivalent of an “M-mode” of the diaphragm on echocardiography), respiratory bellows which are placed around the chest wall, or a signal from the respirator if the patient is under anesthesia. Lines of k-space are continuously acquired

16

M. A. Fogel

during the cardiac cycle and only those lines of k-space which fall within certain positional parameters of the diaphragm or chest wall are incorporated into the image; the others are discarded. Although this is a very inefficient method of imaging, it is very effective and used in imaging coronary arteries, for example, where high resolution is needed. Whole heart angiography is also unsuitable for anything but respiratory gating. 4. Single-shot imaging, where all the lines of k-space are acquired within a single heartbeat. Advances in hardware and parallel imaging have dramatically improved the speed and quality of these single-shot and real-time techniques and are now often used for scanning patients unable to breath-hold. 5. The newer “self-gated” techniques, referenced above, that also compensate for respiration where a 3D cine data set can be acquired at multiple respiratory phases without an ECG or navigator [18].

Another contrast agent that has become increasingly used is ferumoxytol, which is an iron-based contrast agent that is administered slowly over 15 min. This contrast agent has the distinct advantage of having a long half-life in the blood so that high-resolution segmented imaging may occur. This contrast agent is generally used in conjunction with high-­ resolution static inversion recovery gradient echo imaging to visualize coronary arteries, to perform whole heart 3D imaging, or to enhance 4D flow imaging signal [19] (see below). In addition, ferumoxytol has been used to acquire 4D whole heart cine imaging to obtain a 3D beating heart [20], combining anatomy, cine imaging, and 4D flow [19] and for the self-­ gated technique mentioned above that acquires cine at multiple respiratory cycles [18] (see above). The safety profile of these agents is beyond the scope of this chapter. However, it should be noted that there has been a recently published multicenter trial of ferumoxytol demonstrating a safety profile equivalent to other imaging agents and in certain regards, better [21].

1.6 Contrast Agents

1.7 Remaining Motionless in the CMR Scanner: Anesthesia and Sedation

These agents offer another important source of distinguishing tissues from each other besides the intrinsic properties of T1, T2, and T2* for example. The most commonly used imaging agents, the paramagnetic chelates of gadolinium (Gd3+), generally work by predominantly shortening T1 and to a certain extent T2; they generally enhance the signal on T1-­ weighted images. Gadolinium, which has a very large magnetic moment, has unpaired orbital electron spins and shortens T1 by allowing free protons to become bound creating a hydration layer, which helps energy release from excited spins and accelerates the return to equilibrium magnetization. For other contrast agents which predominantly shorten T2, the reverse is true; shortened T2 leads to decreased signal on T2-weighted images. The effects of these agents can be described by the following formulae:



1 1 = + r1C T1 T10 1 1 = + r2 C T2 T20

where T10 and T20 are the relaxation times prior to and T1 and T2 are the relaxation times after contrast agent administration, C is the concentration of the agent, and r1 and r2 are the longitudinal and transverse “relaxivities” of the individual agent (which are field strength dependent). CMR applications which utilize these agents include delayed enhancement, first pass perfusion, coronary angiography in certain sequences, and characterization of tumors and masses.

The degree of cooperation necessary for successful performance of CMR is generally greater than that of any other type of MRI examination; scans require no significant movement, repeated breath-holds at the same point of the respiratory cycle over a period of 45  min to an hour, and can be lengthy. Couple this with the strange environment of the scanning room and the loud banging noises, it is no wonder that both adults and children alike find this very intimidating. Therefore, the use of medication may be required; either conscious sedation or general anesthesia is generally administered so that children who are too young to cooperate or adults with congenital heart disease who may not want to cooperate for one reason or another (e.g., claustrophobia) can still undergo successful CMR. With conscious sedation, patients continue to breathe throughout the scan and imaging has to be substantially altered because of this whereas in a paralyzed, intubated, and mechanically ventilated patient under general anesthesia, the effect of “breath-holding” can be created by having the anesthesiologist temporarily suspend ventilation. This is not to say that anytime a patient undergoes general anesthesia in the CMR environment that suspending respiration should be performed but rather that this technique is available to the CMR imager. It should be noted that imaging using sedation or general anesthesia with free breathing is much more physiologic than imaging with positive pressure mechanical ventilation and breath-holding (see above) and, therefore, may be more advantageous than the minor increase in image fidelity with breath-holding. For example, a single ventricle patient after Fontan depends

1  General Principles of Cardiac Magnetic Resonance Imaging

upon both cardiac and respiratory effects to allow for pulmonary blood flow; suspending respiration may alter the physiology artificially and therefore, although accurate for suspended respiration, the physiology is not reflective of the patient’s true state. In addition, because systemic venous return changes during the respiratory cycle, imaging during suspended respiration will obtain data only in that state while if the patient is imaged during free breathing, the loading conditions across the respiratory cycle is “averaged” into the image and is more reflective of the patient’s true physiologic state. There is no definitive cut-off age for the age range where medication is needed to remain motionless for a successful CMR study; however, in general, most children greater than or equal to 10–12 years old can cooperate. Of course, this is just a rule of thumb as there can be 7-year-olds who are very mature and can follow directions while there are some 15-year-olds who will simply not cooperate and will require pharmacology. Limited scans with reduced times may be possible with younger patients who would normally require conscious sedation or general anesthesia and this may be considered; it is all in the judgment and purview of the family, physicians, and other healthcare providers caring for the patient. Preparation of the child prior to the scan is important; the involvement of child life experts, a supportive parent or other regular caregiver in the scanner room can reduce anxiety and be the difference between a scan under medication, without medication, or a successful versus an unsuccessful scan. As the CMR environment can be a challenging one for the anesthesiologist or the pediatrician/nurse sedation specialist, monitoring is extremely important since the patient’s body will be mostly within the scanner itself; direct visualization during the study is not possible without removing the patient from the bore of the magnet and removing the coil. Many centers utilize a direct video feed with cameras designed to work within the CMR environment and placed in critical positions. For example, a camera pointed down the bore of the magnet is essential along with cameras in other areas to get a good view of what is occurring in the scan room. In addition, extensive physiological monitoring of subjects using equipment specifically designed to be operated in the MR scan room is essential for the safe conduct of the study. Pulse oximetry, limb-lead ECG, blood pressure monitoring, inspiratory and expiratory gas analysis such as end-tidal carbon dioxide, and temperature monitoring (especially in young children) should all be available and used. The monitoring systems should be available wherever the anesthesiology/sedation teams are positioned; this is generally either in the control room or scan rooms. Many facilities position the anesthetic equipment and gas tanks directly outside the scan room, with the gas lines passing through “wave guides” in the wall of the scanner room installed for just this

17

purpose. This arrangement has two advantages: (a) there is reduced risk of inadvertently introducing non-CMR compatible equipment into the scan room and (b) communication between the anesthesiology/sedation team and imaging teams is much easier in this setup. It should be noted, however, that this comes at the cost of increased compliance in the anesthetic circuit. If the decision is made to keep monitoring and anesthetic equipment in the scan room, there is usually a minimum distance that this equipment must be kept from the magnet within which it may not operate correctly, may interfere with the images, and might even be attracted into the scanner bore. Careful establishment of this distance from the manufacturer is mandatory before the equipment is first introduced into the scan room. Even the use of physical restraints to prevent incursion of the equipment within such a distance, and thus avoid accidents, should be considered. Direct verbal communication between the anesthesiology/sedation teams and the imaging teams should be on-going at all times with visual contact preferably as well. Neonates and very small infants less than 6 months of age may undergo CMR successfully while sleeping using a “feed and swaddle” technique [22, 23]. The patient usually is kept awake for a while prior to scanning (3–4 h); when the child enters the preparation area, the intravenous is inserted and the ECG leads are placed. At this point, the baby is very fussy; however, feeding the infant and subsequently swaddling with a warm blanket in a quiet and dimly-lit environment prior to the study will allow the patient to fall asleep; the patient is subsequently transported to the scanner room. Vacuum-shaped support bags can also be utilized to reduce patient motion; placing ear plugs, a hat over the head and ears as well as blankets around the head all aid in keeping the child comfortable and asleep. Imaging sequences that allow for free breathing must be used. Whether to use deep sedation or anesthesia to perform CMR has been debated for many years. Consideration should be given to how long the CMR scan is likely to take, the patient’s age, the flexibility of CMR scanner time, and the availability of anesthesiology staffing and/or the availability of specialized sedation teams which include nurses and pediatricians. The practice is obviously a matter for individual, institutional, and patient preferences. Anesthesia is much more predictable when it comes to onset of action and duration/depth of impaired consciousness; this is advantageous in scheduling CMR examinations and running the schedule smoothly. Deep sedation use has been associated with reduced image quality in some studies [24] but not in others [25], and in some institutions, is far more likely to fail than anesthesia [26], though failure rates can be reduced to close to zero [25] by careful use of expert personnel and strict sedation regimes [25, 27–30]. Imaging performed under anesthesia can be shorter “in theory” because of the

M. A. Fogel

18

ability to breath-hold; in practice, however, the scanning time difference is marginal at best and breath-holding, as mentioned above, is less physiologic. Anesthesia has been reported to be marginally safer than deep sedation in some studies [24, 31] and equal in others [25], but there is no doubt that it is more costly and invasive. There are numerous pediatric centers with many years of experience at performing CMR under deep sedation with excellent safety records [25, 27, 28, 30]. The end result is that both techniques are likely to remain in practice for the foreseeable future.

1.8 The Standard Pediatric/Congenital Heart Disease Examination There are numerous protocols for a standard CMR examination of the heart, many equally as valid as the other. The one presented in this chapter is meant to be as complete and as efficient as possible; however, it should be recognized that this is not the only one. As each phase of the protocol is delineated, the technique utilized will be expanded upon in detail to give the basics of the different types of CMR.

1.8.1 Axial Imaging (Fig. 1.18) The initial part of the examination begins with a set of static steady-state free precession (bright blood) images in the axial (transverse plane) extending from the thoracic inlet to the diaphragm. Generally, 45–50 contiguous end-diastolic slices are obtained of three (for babies) to 5 mm in thickness; end-diastole is acquired by placing a “delay” after the R wave of the ECG. At this point in the cardiac cycle, the heart is relatively motionless, allowing for high-fidelity imaging. This set of data, which usually takes two and a half to four and a half minutes to acquire (depending upon the patient’s heart rate and size), is utilized as a general survey of the anatomy and may be used as a localizer for subsequent higher fidelity cine imaging, flow measurements, etc. These images are usually acquired with multiple averages (generally 3) during free breathing. In babies, to maintain signal to noise but nevertheless obtain thinner slices, overlapping slices can be used; the cost is prolonged acquisition time. 1. From this survey, a number of features may be gleaned with regard to cardiovascular structure in congenital heart disease [32]: (1) the position of the heart in the chest and IV

liver

I

LPA

RPA Sp

II

RV

V

RSVC LV CS

Atria

LSVC Az

Az III

LVOT

RSVC

LSVC

VI

LV DAo

TAo

Az

Fig. 1.18  Selected initial axial images of a patient with heterotaxy and complete common atrioventricular canal. Note how much can be gleaned from the first set of static steady-state free precession images. Images progress from inferior to superior as the roman numerals increase from top to bottom and from left to right. In I, a transverse abdominal view shows a midline liver and spleen (sp) on the right. In II, note the complete common atrioventricular canal, the dilated coronary sinus (CS),

and dilated mildline azygous (Az). In III, note the widely patent left ventricular outflow tract. IV (top right) demonstrates the main pulmonary artery as well as the right (RPA) and left pulmonary artery (LPA) being confluent. In V, note how the dilated AZ enters the right superior vena cava (RSVC) as well as the presence of a left superior vena cava (LSVC). Finally, in VI, note the left aortic arch along with the RSVC and LSVC without a bridging vein. TAo transverse aortic arch

1  General Principles of Cardiac Magnetic Resonance Imaging

in which direction the apex is pointing, (2) normal cardiac segments (atria/ventricles/great arteries), (3) the intersegmental connections (atrio-ventricular and ventriculo-­arterial), (4) veno-atrial connections, (5) aortic arch anatomy such as coarctation of the aorta and sidedness of the aortic arch, (6) pulmonary arterial tree (such as pulmonary stenosis, pulmonary sling), (7) extracardiac anatomy and its relationship with the cardiovascular ­system such as the trachea and tracheobronchial tree, abdominal situs such as the position of the liver, spleen, and stomach, qualitative assessment of lung size (e.g., important in Scimitar syndrome). For lesions such as main and branch pulmonary artery stenosis and coarctation of the aorta, off-axis imaging planes are necessary to confirm and better display these findings; however, these lesions can often be inferred from the stack of axial images. Qualitative assessment of lung hypoplasia and unbalanced pulmonary blood flow can be roughly estimated by the pulmonary vascular markings. If the study is ordered to determine if the patient has a vascular ring, the diagnosis is nearly always readily obtained using this stack of images using the feed and swaddle technique if an infant mentioned above [23]. Image acquisition time for the three-dimensional dataset can be accomplished in 20–30 s depending on the patient’s heart rate. 2. There are drawbacks to using the axial stack when using it for diagnostic purposes; it must be remembered that is solely a survey and to be used as localizers for higher resolution imaging. As examples, smaller anatomic structures such as the pulmonary veins may not be visualized well or seem to appear to be connected anomalously but really be connected normally because of partial volume effects. Follow-up with off-axis imaging is mandatory. 3. In addition to the axial stack of SSFP images, a set of HASTE (Half-Fourier-Acquired Single-Shot Turbo Spin Echo) axial images (Fig. 1.19) can be very useful and are usually obtained while multiplanar reconstruction is being performed on the SSFP images (see below). HASTE is a dark blood, single-shot (image obtained in one heartbeat) technique which is low resolution and acquired during free breathing, generally obtained in 1–2 min. If the RR interval of the patient is under 600 ms, the images are generally acquired every other heartbeat (doubling the acquisition time) to allow the protons to relax further. HASTE images are less susceptible to flow artifacts and metal artifacts. For example, turbulence in the systemic to pulmonary artery shunt (Blalock-Taussig shunt) or Sano shunt in a single ventricle patient after Stage I Norwood reconstruction will demonstrate signal loss in the shunt itself and the pulmonary arteries on SSFP imaging. Turbulent flow occurs in diastole as well as systole in this scenario and recalling that the static SSFP images are acquired in diastole, these structures are

19

difficult if not impossible to see on the SSFP images. These structures are, however, readily visualized on the HASTE images. Multiple patients can present with braces on their teeth which is common in adolescents as well as stents in their great arteries or other blood vessels; these metallic objects can and generally do produce artifacts which appear on the SSFP imaging, but not on the HASTE images. Note, however, that because of the “cage effect” (see below in the dark blood section 1.8.3), direct measurement of the cavity of the stent is not possible. HASTE imaging can also be useful with visualizing regions of the coronaries and in characterizing masses; however, dedicated subsequent imaging of these structures is mandatory. The HASTE images give the imager a “first pass” at the problem and, similar to the stack of SSFP imaging, is simply a survey.

1.8.2 Multiplanar Reconstruction During the acquisition of the HASTE images, multiplanar reconstruction is performed on the axial SSFP (or HASTE) images. Multiplanar reconstruction is the act of taking the contiguous stack of images and reconstructing these images into other planes (e.g., axial images being resliced as coronal images or in a double oblique angle to obtain the “candy cane” view of the aorta). Nearly all scanners today come with software which allows this to be readily performed. The purpose of this obviously is to obtain orientation and slice positions for dedicated images of the anatomy in question, functional imaging, tissue characterization, and blood flow. Further anatomy can be obtained with cine, the various types of dark blood imaging, or 3-dimensional contrast-enhanced images (see below). For the 3-dimensional contrast-enhanced slab, these axial images act to ensure that the anatomy in question is covered by the slab. Ventricular function and blood flow are obtained using cine and phase contrast magnetic resonance (PCMR) (see below). Off-axis imaging planes can be used, for example, to profile the ventricular outflow tracts, the atrio-ventricular valves, major systemic and pulmonary arteries and veins and all their connections to the heart.

1.8.3 Dark Blood Imaging (Figs. 1.8 and 1.20) High-resolution dark blood imaging (as compared to the low-resolution HASTE images) is static in nature and is used sparingly because it is time consuming; 1–2 images can be obtained in a breath-hold. There are numerous types of dark blood imaging such as T1 weighting, T2 weighting, spin echo imaging, turbo spin echo imaging, double or triple inversion recovery, etc. This technique is generally utilized for tissue

M. A. Fogel

20 HASTE Images

RV LV

MPA

LPA

RPA

Fig. 1.19  Two initial axial HASTE images of the patient in Fig. 1.18 with heterotaxy and complete common atrioventricular canal. The upper and lower panels are equivalent to panel II and IV in Fig. 1.18; compare these images with those of Fig. 1.18

characterization and to define anatomy when turbulence or artifacts get in the way of bright blood techniques. The blood from the heart cavities and blood vessels is black while soft tissue is signal intense. Most dark blood imaging in children utilizes either T1- or T2-weighted imaging with the double inversion approach. The details of how each type of dark blood imaging is created are beyond the scope of this chapter; however, a simple example is instructive. The double inversion T1-weighted dark blood technique is utilized to maximally suppress signal from blood and begins with a nonselective inversion pulse which can be thought of as flipping all the protons 180° throughout the body, destroying all the signal from these spins. This is subsequently followed by a selective inversion pulse which flips the protons once again 180° but in a selected region of the body (such as the imaging plane needed); a standard T1-weighted spin echo sequence is then run. In this way, all the blood entering the imaging plane is signal poor with the spins destroyed in the nonselective inversion pulse and detailed endocardial or endovascular surfaces can be visualized. Dark blood imaging can be used, MPA

MPA

PDA

PDA

TAo C

RV

DAo

DAo TAo SCA

PDA TAo DAo

Fig. 1.20  Dark blood, 3-dimensional gadolinium and “fly-through” imaging of a neonate with hypoplastic left heart syndrome who has not undergone surgery. The left upper image is an off-axis sagittal view demonstrating the right ventricular outflow tract giving rise to the main pulmonary artery (MPA), patent ductus arteriosus (PDA) connecting to the descending aorta (DAo). The upper middle is similar to the upper left except a few millimeters over to the right demonstrating the hypo-

plastic transverse aortic arch (TAo), the coarctation (C), and the DAo. The right upper and right lower images are 3-dimensional reconstructions from a time-resolved gadolinium sequence which demonstrates the MPA, PDA, hypoplastic TAo and DAo from a sagittal (top) and posterior (bottom) view. The lower left is a “fly-through” image of the 3-dimensional reconstruction looking up from the DAo towards the os of the PDA, hypoplastic TAo and subclavian artery (SCA)

1  General Principles of Cardiac Magnetic Resonance Imaging

as mentioned above, to characterize different types of tissue as these will generate different signals. As an example, fat will be intensely bright on T1-weighted imaging while myocardium will be much less so. In addition, special pulses can be used to change the signal intensity and determine if indeed this tissue is what is suggested; taking for example fat as was just mentioned, a “fat saturation” pulse may be coupled with dark blood imaging and will turn the very bright signal of fat without this fat saturation pulse into a very dark signal with the fat saturation pulse, confirming that the bright signal is indeed fat. This may be useful for lipomas—visualizing this mass on T1-weighted images with and without a fat saturation can confirm the diagnosis. Triple inversion recovery may be used to delineate edema in the tissue from, for example, a myocardial infarction or myocarditis. Typically, in our imaging protocols, if needed, this is performed after the SSFP and HASTE imaging but should not be used after gadolinium administration except for specific applications such as myocarditis or tumor characterization. If used after gadolinium administration, the blood pool will demonstrate signal which is counterproductive to the intent of dark blood imaging in the first place. Use for dark blood imaging besides myocarditis is to visualize the pericardium, image the tracheobronchial tree (useful in a vascular ring study), and tumor characterization (with and without gadolinium). As mentioned above, it is useful to image patients when coils, stents, braces, spinal rods, and other foreign material cause artifacts on bright blood imaging. Precise measurements cannot be performed within a stent, however, because of the “cage effect.” The image artifact caused by the stent prevents the physician from seeing the critical area in and around the stent. This is caused by the fact that a metallic stent behaves as a “Faraday Cage” due to its geometry and material, and the stent additionally creates a magnetic susceptibility artifact due to the material of manufacture of the stent. A more modern approach is the use of a 3D dark blood acquisitions such as SPACE (Sampling Perfection with Application-optimized Contrasts using different flip angle Evolution) which is not only gated to the cardiac cycle but also employs a navigator pulse to obtain high-resolution (1–1.2 mm isotropic) imaging. It is generally performed in systole, aiding to null the signal from blood. It has found utility in patients who have hardware in place such as stents or devices or where turbulence makes SSFP imaging problematic.

1.8.4 Cine (Figs. 1.10, 1.13, 1.15, 1.16, and 1.17) Myocardial motion and blood flow can be visualized with cine imaging to determine function and physiology. It is one of the two workhorses of CMR in this regard (the other being

21

phase contrast velocity mapping which will be discussed next). The two major types of cine imaging are unbalanced gradient echo imaging and SSFP as mentioned above. The unbalanced gradient echo technique is older but still has a number of uses in state-of-the-art CMR. For example, unbalanced gradient echo imaging is useful to determine valve morphology using a high flip angle (Fig. 1.13); in this way, flowing blood into the imaging plane is bright and outlines the leaflets of the valve very well. It is also useful when artifacts plague SSFP as a low echo time (TE) and high bandwidth gradient echo image is less susceptible to these artifacts. High TE gradient echo imaging will enhance turbulence (where SSFP is less susceptible to turbulence) and may be useful in identifying these areas of flow disturbances. It is also the cine workhorse when ferumoxytol is administered as the contrast agent, using a high bandwidth. High-resolution SSFP cine imaging (Figs. 1.15, 1.16, and 1.17) can demonstrate exquisite images of the myocardium, valves, blood pool, and vessels. In assessing myocardial function, it is the technique of choice and the gold standard. These cine images provide excellent spatial and temporal resolution for the assessment of global and regional myocardial wall motion. These cine sequences should be retrospectively gated so that wall motion data is available for the entire cardiac cycle. As mentioned, with prospective triggering, the phases prior to the R wave is generally truncated as noted above in the physics section. With retrospective gating, the number of calculated phases should be figured so that there is only one or less interpolated phase between each measured phase; an interpolated phase is one that shares data between the two measured phases. This is easily performed by doubling the patient’s RR interval and then dividing by the heartbeat “TR” (line TR  ×  number of segments) to get the maximum number of phases or dividing by the number of phases desired to obtain the maximum TR needed. Temporal resolution should be set to provide, in general, 20–30 phases across the cardiac cycle, depending upon the heart rate. Obviously, in a patient with a heart rate of 150 beats/min (R-R of 400 ms), 20 frames/heartbeat is more than adequate (20 ms temporal resolution) while if the heart rate is 50 beats/min (R-R of 1200 ms), 20 frames/heartbeat is not sufficient (60  ms temporal resolution). This is because systole does not vary too much as a function of heart rate; it is diastole that lengthens or shortens. A 60 ms temporal resolution for a heart rate of 50 beats/min will not capture enough frames in systole to adequately assess the ventricle. When an entire ventricular volume dataset is acquired, ventricular volume and mass at end-diastole and end-systole are measured yielding end-diastolic and end-systolic volumes, stroke volume, ejection fraction, cardiac output, and mass [33–37]. To perform an entire ventricular volume set, generally a 4-chamber view is first obtained by cine (orientation and slice position determined by multiplanar reconstruc-

22

tion as noted above); the 4-chamber view is defined as the view that intersects the middle of both atrioventricular valves at the atrioventricular valve plane and the apex of the heart. Subsequently, a series of short axis views are obtained which are perpendicular to the 4-chamber view and span from atrioventricular valve to apex. It should be noted that this requires obtaining short axis slices one slice past the atrioventricular valve level and one slice past the apex to ensure the entire volume is obtained; this can be clearly positioned off the 4-chamber view at end-diastole. Measurement of ventricular volumes involves contouring the endocardial border of each slice of a given phase (e.g., end-diastole or end-systole) from base to apex and planimeterizing this area. The product of the sum of the areas on each slice encompassing the ventricle and the slice thickness yields the ventricular volume at that phase. This procedure, if performed at end-diastole and end-­ systole, will yield two values and the difference between these values is the stroke volume; and the ratio of the stroke volume to the end-diastolic volume multiplied by 100 yields the ejection fraction. The cardiac output is obtained by multiplying the ventricular stroke volume and the average heart rate during the cine acquisition (note that if there is atrioventricular valve insufficiency, a ventricular septal defect or there is double outlet ventricle, this will not equate to aortic outflow). Ventricular mass is similarly measured, generally at end-diastole, by contouring the epicardial border on each slice which contains ventricle and planimeterizing the area which contains both the ventricular volume and mass. This value is subtracted from the ventricular volume measurement at each slice and yields the ventricular mass multiplied by the density of the myocardium. Most scanners come with and numerous independent companies sell software which semiautomates this process; ventricular volumes and mass can generally be obtained in a few minutes of post-­processing. More tedious is contouring ventricular volumes through every phase of the cardiac cycle; however, this will yield a ventricular volume–time curve which may be useful in some situations. Because CMR can acquire multiple contiguous, parallel tomographic slices, there is no need for geometric assumptions, making the technique an excellent tool for precise measurement of ventricular volumes and mass in congenital heart disease. Indeed, cine CMR is considered the “gold standard” for ventricular volume and mass in both adult cardiology, pediatric CMR, and congenital heart disease. Ventricular size and shape can vary considerably in various forms of congenital heart disease (e.g., single ventricles, corrected transposition of the great arteries, etc.). Ventricular cine is also utilized not only for global but for regional wall motion abnormalities as well. Besides ventricular size, mass, and wall motion, cine imaging is excellent for identifying vessel sizes as well as stenosis or hypoplasia including great arteries, along a ventricular outflow tract or in a baffle or conduit. On the flip

M. A. Fogel

side, cine can also be used to qualitatively assess regurgitation of valves of which there should be minimal in the normal heart. All this is determined not only by the shape of the vessel but by loss of signal due to acceleration of flow through a stenotic vessel/valve or a regurgitant valve; a classic example is the flow void through coarctation of the aorta. In addition, shunts can be identified by cine as turbulence visualized across atrial or ventricular septae will indicate interatrial and interventricular communication. A way in which shunting can be accentuated visually is with the use of a presaturation tag. When the protons in a plane of tissue are flipped 180° to destroy their spins prior to imaging (similar to a selective inversion pulse in a plane of tissue intersecting the imaging plane), a presaturation tag is said to be laid down. This presaturation tag labels the tissue it intersects with decreased signal intensity (black on the image). If the presaturation tag is laid down on the left atrium prior to a gradient echo sequence in a patient with an atrial septal defect, blood flowing from left to right will be dark on the bright blood cine and visualized to cross from left to right atria. Similarly, if there is right to left flow, bright signal from the right atrium would be seen to cross to the darkness of the left atrium. By stringing a series of continuous single-shot images together in one plane (“single plane, multiphase” imaging), motion can be captured and this is termed “real-time cine CMR” (see above). Essentially, the SSFP technique is used to acquire all the lines of k-space needed to create an image continuously in the same plane. “Interactive real-time cine CMR” adds the ability to be able to manipulate the real-time imaging plane interactively, similar to echocardiographic (“sweeps”); this provides a fast way to assess cardiovascular anatomy, function, and flow. These images can be used for localization for higher resolution regular cine CMR and have been utilized in the past to actually acquire fetal cardiac motion. It is also used in the event there is too much arrhythmia so that at least a qualitative assessment of the heart can be made. Temporal resolution can be as low as 35 ms using parallel imaging. CMR techniques in general and cine in particular build the image of multiple heartbeats. If multiple averages (excitations) are used, this can be in the hundreds. The disadvantage to this is the time it takes to acquire the data unlike “real-time” CMR cine imaging just mentioned or echocardiography where the cardiac motion is instantaneously obtained. The distinct advantage to this approach, however, is because the image is built over many heartbeats; it represents the “average” of all those heartbeats over the acquisition time. This truly is an advantage as it would be assumed that this “average,” embedded in the image, is more reflective of the patient’s true physiologic state than the instantaneous images of echocardiography. To perform the equivalent, the echocardiographer would have to view hun-

23

1  General Principles of Cardiac Magnetic Resonance Imaging

dreds of heartbeats and “average” it in the imager’s mind with all the subjectivity that entails. Picture measuring hundreds of M-mode dimensions of the ventricle (end-diastolic and end-systolic diameters) and then averaging them all together to get a shortening fraction: that is what is obtained with cine CMR in one image when measuring ventricular function parameters!

1.8.5 Phase Contrast (Encoded) Magnetic Resonance (PCMR) (Figs. 1.13, 1.14, 1.21, 1.22, and 1.23) [38–46] This CMR technique, also known as “velocity mapping,” is used to measure flow and velocity in any blood vessel with few limitations (for example, generally 4–6 pixels must fit across the blood vessel in cross-section for it to be accurate). Broadly speaking, there are two types of PCMR— through plane which encodes velocity into and out of the imaging plane and in plane which encodes velocity in the imaging plane (as in Doppler echocardiography). For example, through plane PCMR can measure cardiac output, the

LPA

RPA

aortic, and the pulmonary valves by measuring the velocities across the valves in cross-section, multiplying by the pixel size summed across the entire vessel, integrated over the entire cardiac cycle, and multiplied again by the heart rate. In the absence of intracardiac shunts, the flows across the aortic and pulmonary valves should be equal—an internal check to the measurement. In another example, relative flows to each lung may be measured by utilizing throughplane velocity maps across the cross-section of the right and left pulmonary arteries, obviating the need for nuclear medicine scans. By placing a through-plane velocity map across the main pulmonary artery, an internal check on the branch pulmonary artery flows is obtained as the sum of the blood flow to the branch pulmonary arteries must equal the blood flow in the main pulmonary artery. In addition, it is also common to utilize PCMR as a check on cine measurements (e.g., cardiac index of the aorta should be equal to the cardiac index of the left ventricle in the absence of mitral insufficiency or intracardiac shunts). It is clear that this is a strength of CMR—the ability to perform these checks for internal validation of the quantitative data, unlike other imaging modalities.

LPA

RPA

RAA AAo PAB PAB

MPA LV

LV

RV

LV

LPA RPA

PAB

MPA

PAB

MPA

DAo LV

Fig. 1.21  Various types of imaging in an infant with corrected transposition of the great arteries after a pulmonary artery band (PAB). The upper left and middle panels are two views of a 3-dimensional gadolinium image of the right-sided circulation showing the left ventricle, main pulmonary artery (MPA), PAB, and the right (RPA) and left pulmonary arteries (LPA) from anterior (left) and anterior tipped up to transverse view (middle). Note how the right atrial appendage (RAA) is easily seen. The right upper and left lower images are 3-dimensional gadolinium reconstructions of both circulations demonstrating the ante-

rior aorta and branch pulmonary arteries form the anterior (upper right) and posterior (lower left) views. The lower panel second from the left and second from the right are magnitude and in-plane phase images from phase-encoded velocity mapping demonstrating the left ventricular outflow tract and showing the jet through the PAB (signal intense is caudad) with a VENC of 400 cm/s. The right lower image is an orthogonal view through the left ventricular outflow tract demonstrating the turbulence distal to the PAB

M. A. Fogel

24 Fig. 1.22  Data and images from a patient with a bicuspid aortic valve, aortic stenosis, and insufficiency. The graph of flow versus time is on the lower left and on the upper left is the relevant data. Gradient echo images of the left ventricular outflow tract in two orthogonal views demonstrating the aortic insufficiency jet during diastole (arrow). AAo ascending aorta, LV left ventricle

Aortic Regurgitation Cardiac output = 3.8 l/min/m2 Forward flow = 108 cc Reverse flow = 38 cc Regurgitant fraction = 35% Heart rate = 87 BPM Peak velocity = 3.7 m/s

AAo

LV

AAo

LV

Fig. 1.23  Data and images from a 2-year-old with an atrial septal defect (ASD) of the inferior vena cava type and anomalous right pulmonary venous connections to the right atrium (RA). The off-axis sagittal magnitude image (upper right) demonstrates the ASD while the in-plane, colorized phase-encoded velocity map (lower right) in the same orientation as the magnitude image demonstrates left to right flow by the red color jet as in echocardiography (red is caudad and blue is cephalad flow). The aortic and pulmonary flow are both graphed simultaneously (lower left); data demonstrates as Qp/Qs 2.3

Aortic flow Cardiac output = 4.6 l/min/m2 Pulmonary flow Cardiac output = 10.7 l/min/m2 Qp/Qs = 2.3

ASD

Pulmonary

Aorta

RA

1  General Principles of Cardiac Magnetic Resonance Imaging

It is key to understand what the “phase” means in the term “phase-encoded velocity mapping” for one to understand how this is used to measure flow. Phase was discussed in the physics section and will be explained in a slightly different way in this section, although it represents the same physical principle. When tissue is excited by radiofrequency energy, the subsequent signal that gets generated when the protons relax (for example, a sine wave) can be described by its frequency (how many cycles per second), its amplitude (the strength of the signal), and its phase (where, in a given time, is the sine wave in its cycle). Two waves can have the same frequency and amplitude but be in different points in their cycle (i.e., they are identical waves but shifted in time); they are out of phase. Think of two identical sine waves placed one atop the other in a signal amplitude-time graph (signal amplitude on the Y axis and time on the X axis) and then shift one slightly to the right in time; these sine waves are identical but out of phase with each other. Another way to understand this is that the same part of each of the two identical sine wave occurs at a different point of time (e.g., the peak of sine wave “A” occurs prior to the peak of sine wave “B”). The principle underlying phase contrast CMR very simply is that moving tissue within a magnetic field that has a gradient applied to it changes phase. Put another way, whenever anything moves along the axis of an applied gradient, the phase of the spinning vectors in that object becomes altered relative to the stationary object. Remember the Larmor equation states the precessional frequency of the protons is directly proportional to the magnetic field; if that magnetic field is altered by position creating a gradient, the precessional frequency will change depending upon position. Any tissue moving across the gradient will change precessional frequency and accumulate phase shift selectively “labeling” itself. Phase contrast velocity mapping utilizes a “bipolar” radiofrequency pulse which is equal in magnitude but opposite in direction (e.g., turns from positive to negative and then from negative to positive); this is done with two back-to-back pulses with a slight delay. The sequence has the following effect: Before any radiofrequency pulse is applied, protons are tilted as a function of where they are positioned in the magnetic gradient. When the first radiofrequency pulse is applied, both stationary and moving tissue protons are further tilted; when the second equal and opposite radiofrequency pulse is applied immediately afterwards, protons of the stationary tissue return the tilt of their protons back to their original position and accumulate a net phase of zero (their tilt goes back to their original position since they experienced an equal and opposite radiofrequency pulse and haven’t moved position—they are experiencing the same magnetic field before and after these pulses). Protons of the moving tissue, however, do not revert to their original tilt since they have moved and are experiencing a different magnetic field because of the magnetic gradient in position.

25

These protons are said to have accumulated a “phase shift.” To summarize, this will yield a zero phase change for stationary objects when both pulses are applied whereas there will be a net accumulation of phase in moving tissue. By subtracting, pixel by pixel, the phases of one pulse from the other, background phase changes of stationary objects are canceled out and the phase shift of the moving tissue is amplified. Then, usually, the “phase difference” method is used to map the phase shift angles into signal intensities. Flow is calculated by the formula:

∆ phase = g × v × T × Ag

where g = gyromagnetic ratio, v = velocity, T = duration of the gradient pulse and Ag is the area of each lobe of the gradient pulse. By tailoring the strength of the radiofrequency pulse to the anticipated velocities, accurate measurement can be obtained; this is called a VENC (velocity encoding) and is the equivalent of the Nyquist limit in echocardiography. Using the VENC and the signal intensity, the velocity of moving tissue in each pixel can then be encoded. This can occur with either blood (hence blood phase-encoded velocity mapping) or with myocardial tissue (also called myocardial velocimetry and is the equivalent to Doppler tissue imaging). As mentioned, spatially, there are two ways to encode velocity in the image: (a) “through plane” where each pixel encodes velocity into and out of the plane of the image and (b) “in-plane,” where velocities are encoded in the plane of the image and not into and out of the plane similar to Doppler echocardiography. Unlike Doppler echocardiography, however, velocities are encoded in either the Y- or X-direction of the image. This is advantageous as each pixel can encode velocity in three orthogonal planes. Motion in one direction is mapped onto the anatomic image as increased signal intensity or bright and motion in the other direction appears signal poor and dark; stationary tissue appears gray. Air, such as in the lungs or around the patient, is encoded in a mosaic black and white. Color can be added to make it similar to Doppler echocardiography. Flow measurements by this methodology utilize the cross-sectional area of the vessel perpendicular to the transaxial direction of flow. All software can identify the regions of interest simultaneously on the anatomic or “magnitude” images as well as the “phase” images (which can sometimes be difficult to read). The product of the velocity and the area of an individual pixel give flow through that pixel; summing this across the vessel cross-sectional region will yield the flow at a given phase of the cardiac cycle. Integrating this across the entire cardiac cycle (i.e., each phase plotted as a time-flow curve), the flow during one heartbeat is calculated. The product of the heart rate and this flow in one heartbeat will yield the cardiac output.

26

Through plane phase-encoded velocity mapping can also be used to assess the peak velocity for gradient calculations as might be used for coarctation of the aorta or ventricular pressure estimates as with a tricuspid regurgitant jet. If the plane is perpendicular to flow in the region of maximum velocity, the greatest velocity in any pixel in the region of interest in the cardiac cycle is used in the simplified Bernoulli equation. It should be noted that planes not perpendicular to flow or one at a level where the maximum velocity is not present will underestimate this number; generally, this is used in conjunction with in-plane velocity mapping (see below) to measure maximum velocities. This is similar to Doppler echocardiography where, if the plane is angled obliquely to the jet of interest or the sector is not in the area of maximum velocity, an underestimate of the maximum velocity will occur. Encoding velocity parallel to flow (“in-plane” velocity encoding) is predominantly used to measure peak velocities which is similar to what Doppler echocardiography measures. The advantage of this type of phase encoding over the through plane technique in measuring maximum velocities is that velocities can be measured along a jet of interest in the direction the jet is pointing (similar to continuous wave Doppler), through plane mapping is placed at a certain level of the jet and the velocity measured which is similar to the range-gating technique of pulse wave Doppler echocardiography. With in-plane velocity mapping, the jet is aligned by rotating the entire field of view to make one side exactly perpendicular to the jet. The peak flow velocities can then be translated into pressure gradients via the simplified Bernoulli equation. The phase maps on present-day scanners can give a temporal resolution of about 15–20  ms with non-breath-­ hold techniques. Phase-encoded velocity mapping has limitations. Reliability of both through plane and in-plane velocity mapping is a function of a few factors such as slice thickness (“partial volume” effects may induce inaccuracies in velocity calculations) and the angle of the jet (the jet needs to be aligned perpendicular to the direction of phase encoding, similar in some sense, to Doppler flow measurements). In addition, if the VENC is not chosen close but not below the maximum velocity anticipated, errors may occur. If the VENC chosen is too low, velocities in the vessel will exceed the ability of the CMR scanner to encode them which is akin to aliasing and exceeding the Nyquist limit in Doppler echocardiography. If the VENC chosen is too high, the lower velocities will not be measured as accurately as well as signal to noise decreasing; this is analogous to the difference when measuring a 4  oz of fluid in a 6  oz measuring cup (appropriate setting of the VENC) versus a gallon measuring cup (in appropriate setting of the VENC). Maxwell terms, eddy currents, and whether or not background subtraction is used will also play a role in velocity mapping accuracy.

M. A. Fogel

Finally, by combining through plane and in-plane velocity mapping into the same sequence, CMR can acquire a 3-dimensional slab of velocities termed 4-dimensional or “4D flow” meaning three dimensions of space plus time [47, 48]. Using modern-day gradients and advanced software such as compressed sensing, the entire heart and great vessels can be acquired in anywhere from 3 to 8 min. There are multiple vendors with excellent software that can qualitatively visualize and quantitatively assess these datasets (Fig. 1.14). There are multiple advantages to 4-D flow including being able to “go back” to any blood vessel in the thorax after the patient exam has ended to measure flow; 2D PCMR is aimed at one vessel at a time and once the patient leaves the scanner, there is no one to measure flow in another vessels. In addition, all vessels are measured at the same average heart rate as opposed to 2D flow where it is generally different, making 4D flow more consistent. Further, flow profiles in blood vessels and the heart and their interactions in 4D are easily visualized by this methodology. There are even implementation of these sequences where cine imaging and flow can be acquired at the same time, obviating the need for cine imaging and anatomic imaging (generally with the addition of a contrast agent however) [49]. Finally, metrics such as pulse wave velocity, pressure drop, and energy loss can be calculated whereas with 2-D PCMR, it may be difficult to impossible to do so (see below). There are drawbacks to 4D flow as well. For some implementation of these sequences, administering a contrast agent such as gadolinium or ferumoxytol (see MRA section below) is useful for better signal to noise. PCMR by 2D has many applications in congenital heart disease and in broad categories, they are (a) flow quantification, (b) flow visualization, (c) velocity measurements, and (d) myocardial velocimetry.

1.8.5.1 Flow Quantification 1. Cardiac Output (Figs. 1.22 and 1.23): Measuring cardiac output is an essential factor in assessing cardiovascular performance; this is especially true in patients who have undergone surgical procedures or who have undergone catheter intervention. Lesions such as single ventricles, corrected transposition, tetralogy of Fallot, etc. all benefit from measuring an elementary parameter such as cardiac output. This can also be used as an internal check to the ventricular stroke volume measurements. 2. Regurgitant Lesions (Fig. 1.22): As flow can be quantified as forward and reverse flow, the regurgitant volumes and fractions can be measured and calculated; lesions such as tetralogy of Fallot after repair with a transannular patch [50] or in patients with a bicuspid aortic valve and severe aortic insufficiency all require measurement of

1  General Principles of Cardiac Magnetic Resonance Imaging

leaky semilunar valves. These measurements are readily obtained by placing phase-encoded velocity map at the sinotubular junction (for the aortic valve) and just above the pulmonary valve and measuring the forward and reverse area under the flow-time curve. The regurgitant fraction is simply the area under the reverse flow (regurgitant volume) divided by the area under the forward flow (forward volume) multiplied by 100. An internal check is the measure flow in the cavae which should be the net cardiac output measured by velocity mapping in across the leaky semilunar valve. To obtain how much leakage there is across an atrioventricular valve, a combination of techniques are used; cine CMR is used to measure the stroke volume of the ventricle and phase-encoded velocity mapping to measure the amount of forward flow through the semilunar valve. The difference between the two (with no intracardiac shunts) is the regurgitant volume of the atrioventricular valve. Alternatively, the forward flow across the atrioventricular valve during diastole and the net flow across the semilunar valve in systole can be used (assuming no shunting). 3. Shunts (Fig. 1.23): Many lesions in congenital heart disease have shunting between the circulations present; this shunt flow can be easily calculated by placing velocity maps across the aortic valve and main pulmonary artery and measuring flow (e.g., Qp/Qs) [51]; if there is a aorto-­ pulmonary window, branch pulmonary artery flow is generally used instead. Measuring flow in both branch pulmonary arteries can add an internal check on the amount of pulmonary blood flow for intracardiac shunts and the sum of the flow in the cavae can be used as a check on the amount of systemic blood flow. 4. Flow Distribution to Each Lung: [52] Altered flow distribution to left and right lungs can be common in many lesions in congenital heart disease such as in single ventricle lesions after Fontan, tetralogy of Fallot, or in transposition of the great arteries after arterial switch procedure; all may have branch pulmonary stenosis. Relative flow to each lung is assessed by placing a phase-­ encoded velocity map at each branch pulmonary artery although care must be taken to place the map in the branch pulmonary artery proximal to the takeoff of the first branches to ensure this blood flow is included. Flow measured in the main pulmonary artery must equal the sum of the flows to each lung in the absence of collaterals and is used as an internal check. 5. Collateral Flow: [53] Patients with single ventricles after bidirectional Glenn or Fontan operation develop aorto-­ pulmonary collaterals presumably in response to decreased pulmonary blood flow and cyanosis [54]. In addition, patients with relatively long-standing coarctation of the aorta can develop aortic collaterals which bypass the obstructed segment. In the former, the amount

27

of collateral flow can be quantified in two ways: (1) by measuring flow via velocity mapping across the ventricular outflow tract and subtracting measured caval return and (2) measuring flow in the pulmonary veins and subtracting measured flow in the branch pulmonary arteries. In coarctation, the amount of collateral flow can be determined by placing a phase-encoded velocity map across the aorta just distal to the coarctation site and across the aorta at the level of the diaphragm. Normally, flow at each level will be very similar or the flow at the level of the diaphragm slightly lower (because of flow to the intercostal arteries); however, in the presence of coarctation with collaterals, flow just distal to the coarctation site in the aorta will be lower than flow in the aorta at the level of the diaphragm since collateral flow will present in the latter and not in the former.

1.8.5.2 Velocity Measurements 1. Pressure Gradients: Stenoses of a blood vessel such as a great artery (e.g., coarctation of the aorta, left pulmonary artery stenosis, or a pulmonary artery band, Fig. 1.21) or of a valve (aortic stenosis in a bicuspid aortic valve) can occur in numerous congenital heart lesions. It is important for many reasons to determine pressure gradients in these lesions; as noted above, the determination of gradients by CMR is similar to Doppler echocardiography, using the simplified Bernoulli equation. A maximum velocity is measured, typically in the vena contracta, and the gradient is simply the product of 4 and the velocity (in meter/second) squared. Measurement of maximum velocities may be performed in two ways: (a) “in-plane” velocity mapping directed parallel to the obstruction to flow and (b) “through plane” velocity mapping perpendicular to flow. Both have their strengths and weaknesses (see above discussion).

1.8.5.3 Flow Visualization 1. Septal Defects Using In-Plane Velocity Mapping: Septal defects at both the atrial and ventricular level can be visualized with i2D n-plane velocity mapping. The imaging plane needs to be oriented in the direction of the flow across the defect to successfully visualize it; blood flowing one way would be dark and flow the other way would be bright. In addition, color can be superimposed on the image to simulate color Doppler echocardiography. 2. Flow Directionality in Blood Vessels Using Through Plane Velocity Mapping: A good example of this is isolation or disruption of the subclavian arteries; this may be caused by surgery (e.g., a subclavian flap angioplasty to

28

repair coarctation of the aorta) or may be congenital. When this occurs, the subclavian artery usually is supplied with blood in a retrograde fashion via the vertebral artery; sometimes, with some paraspinal plexuses or other collaterals supply the subclavian artery. Clinically, a “subclavian steal” can occur and CMR velocity mapping techniques can be used to identify retrograde flow in the vertebral arteries. Normally, if an axial through plane velocity map is placed in the neck, both carotid and vertebral arteries will be labeled as either bright or black on the images as they are all flowing in the same direction. With isolation of a subclavian artery, flow in the ­ipsilateral vertebral artery is retrograde while flow in the other three head vessels (carotid and vertebral arteries) will be antegrade. A velocity map in this scenario will encode the vertebral artery ipsilateral to the isolated subclavian artery in one direction (e.g., dark) and the other three head vessels will be encoded in the opposite direction (e.g., bright), proving the physiology. 3. Valve Morphology (Fig.  1.13): Phase contrast velocity mapping is also useful for identifying valve morphology; as through plane phase contrast velocity mapping “tags” flowing blood, this can be used to outline the leaflets tips making a cast of the valve morphology while it opens and closes; the flowing blood is bright or dark and highlights the valve leaflet tips which are gray. Bicuspid, unicuspid, and quadricuspid valves are easily seen. Of particular note, bicuspid aortic valve is very common clinically; it can easily be identified by phase contrast mapping and the degree of valvular stenosis and regurgitation assessed [55, 56].

1.8.5.4 Myocardial Velocimetry This application of phase contrast velocity mapping is the CMR equivalent to Doppler tissue imaging in echocardiography; velocities of the myocardial tissue can be recorded. The phase contrast velocity mapping sequence is configured such that the VENC is set fairly low (15–30  cm/s); modifications in the sequence must be made to keep the TE as low as possible. Doppler tissue imaging can only record myocardial velocities in one direction; that is parallel to the Doppler beam. Myocardial velocimetry, however, is a much more comprehensive measurement of myocardial velocities in that, similar to other phase contrast velocity mapping applications, each pixel can encode velocities in three orthogonal planes; a 3-dimensional velocity map of the myocardium can be measured. Both myocardial velocimetry in CMR and Doppler tissue imaging in echocardiography suffer from the same drawback in that both techniques do not truly measure the velocity of a specific piece of myocardium; they identify a point in space and the velocity of myocardium moving into and out of that point is mea-

M. A. Fogel

sured. Only CMR myocardial tagging and Doppler spectral tracking truly measure the velocity of a piece of myocardium noninvasively. An excellent review comparing the merits of myocardial velocimetry to myocardial tagging was published in 1996. 4D flow imaging has found utility in multiple applications in congenital heart disease [57]. All the above-mentioned metrics with 2D flow can also be calculated and visualized by 4D flow, especially visualizing shunts or regurgitant fraction. In addition, however, pulse wave velocity, which has a bearing on aortic stiffness and ventricular performance, can be obtained easily by multiple 4D flow methodologies [58]. Pressure flow fields and pressure drops across the entire aorta can also be calculated and can come in very useful in lesions such as coarctation of the aorta [59]. In addition, wall shear stress can be derived from 4D flow images in the aorta in patients, for example, with a bicuspid aortic valve [60]. Finally, as a last example, new metrics such as energy loss can be measured and assessed in congenital heart disease, such as in the single ventricle [61] (Fig. 1.14).

1.8.5.5 Magnetic Resonance Angiography (MRA) (Figs. 1.20, 1.21, 1.24, and 1.25) Magnetic resonance angiography, most based on intravenous gadolinium diethylenetriamine pentaacetic acid (Gd-DTPA), can determine detailed anatomy such as major pulmonary and systemic arteries and veins and to glean physiology is a key part of the examination. As discussed above, gadolinium is a paramagnetic element which is administered in a chelated form and is an extracellular agent that changes the magnetic property of the tissue or vessel it is in. It markedly decreases T1 relaxation and which allows its application to distinguish the target structure (e.g., the aorta) from background (e.g., lung, other mediastinal contents). It is considered a highly safe substance with adverse events occurring in one in 200,000–400,000. Usually a “double dose” of contrast is given (for most agents, this is 0.4 cm3/kg; please check the labeling of your individual agent). In the past, much attention has been given to the incidence of nephrogenic systemic fibrosis (NSF) in patients with chronic, severe renal failure, first described in 2000  in 15 patients undergoing hemodialysis who presented with “scleromyxoedema-like” skin lesions. A detailed discussion of this entity is beyond the scope of this chapter; suffice it to say that after modifications of gadolinium use, NSF is nearly eradicated. Very few reports of children developing NSF exist and none under 6 years of age. In addition, there have also been reports of linear gadolinium contrast agents accumulating in small amounts in the brain of individuals exposed to more than four doses of the agent. The recent FDA communication on this issue states that although more testing is needed, data supports the notion that no adverse effects are noted from this issue.

29

1  General Principles of Cardiac Magnetic Resonance Imaging SVC

RPA

I

Ao

LPA

II

III

Ao

IV

V

RPVs

Fig. 1.24  Three-dimensional time-resolved gadolinium imaging in a coronal view. The images are maximum intensity projections (MIPs) and temporally follow the roman numeral from left to right and from top to bottom. Each phase can be created into a 3-dimensional image. Spatial resolution is an isotropic 1 × 1 × 1 mm and temporal resolution is 2.2 s. In I, flow into the superior vena cava (SVC) is seen followed by flow into the right side of the heart (II). In III, flow is now seen in the

Fig. 1.25 Three-dimensional right heart reconstructions of the patient in Fig. 1.24. The right ventricle (RV), main (MPA), right (RPA), and left pulmonary artery (LPA) are clearly seen from an anterior (left), posterior (middle), and transverse (right) view. The RV was removed from the transverse view (right) to facilitate visualizing the branch pulmonary arteries

RPA

LPA

RV

This can be performed anytime during the examination; however, it is performed after cine imaging. It can be performed before or after PCMR; however, if performed before PCMR, the contrast agent boosts the signal of the PCMR flow sequence, giving better signal to noise and more robust

pulmonary veins and the beginnings of the aorta (Ao). In IV, flow has left the right side of the heart and is now mostly in the pulmonary veins, the left side of the heart, and the aorta. In V, flow is seen returning to the heart from the systemic veins. Note the interrupted right subclavian artery in III and IV (arrows). LPA left pulmonary artery, RPA right pulmonary artery, RPVs right pulmonary veins

LPA

RPA

RV

RPA

LPA

MPA

data. Since approximately 10 min is needed between gadolinium administration and imaging for delayed enhancement and even more for T1 mapping (see below), the time after gadolinium administration but before viability imaging is used to perform either phase contrast velocity mapping or

30

3-dimensional static inversion recovery gradient echo imaging (see below). There are two types of contrast-enhanced techniques commonly used to evaluate 3-dimensional anatomy: 1. Static 3-dimensional imaging can be utilized to create a 3-dimension image of the cardiovascular system which can be rotated and cut in any plane desired. The 3-­dimensional slab (see Physics section) can be acquired in any orientation and can be viewed in its raw data format, as a maximum intensity projection, a shaded surface display, or as a volume rendered object. As it is a 3-­dimensional acquisition with one frequency encoding and two phase encoding directions, the thickness of the slices can be much thinner than other techniques—on the order of ¾ to 1 mm. As the resolution is higher than other techniques in CMR, this type of imaging is used to visualize smaller vessels (e.g., aortic-pulmonary collaterals in pulmonary atresia/intact ventricular septum, aortic collaterals in coarctation, etc.). Multiple 3-dimensional data sets are generally obtained which (a) can separate out the systemic and pulmonary circulations and (b) increases the chances of successfully imaging what is needed. Older techniques relied on knowing exactly when the contrast agent reaches the target vessel to begin the 3-dimensional sequence. This was performed in either two ways: (a) Bolus tracking, where the target structure (e.g., the pulmonary arteries in the case of tetralogy of Fallot) is imaged with “real-time” sequence during gadolinium injection—once the contrast agent is seen to arrive at the target structure, the 3-dimensional sequence is then obtained. (b) Timing bolus, where a small amount of contrast agent is injected and timed to determine when it will arrive at the target vessel. Injecting the full dose of gadolinium followed by the 3-dimensional sequence with a delay placed based on the small amount of contrast agent initially given is then performed knowing the gadolinium will arrive at the target vessel based on the initial timing bolus. Modifications need to be made, depending upon the 3-dimensional sequence, if it is “center weighted k-space” or “frontloaded k-space.” The current state of the art is performing a static 3D sequence (inversion recovery gradient echo imaging with ECG gating utlizing the navigator technique and respiratory motion adaptation) either immediately after timeresolved imaging (see below) or with the “slow-drip” of gadolinium technique (which finishes administering gadolinium in the first 1/3 of the sequence) [62]. The navigator technique uses a coil to monitor the diaphragm and allows the algorithm to accept all the data within a certain range of where the diaphragm is (2–4  mm), adjust as

M. A. Fogel

needed, and discard the rest. In this way respiratory motion is compensated for. This 3D technique is generally done isotropic and can be performed with resolutions as low at 0.6 mm using ferumoxytol (see below). Newer techniques under study now allow for creating 3D images without the need for a navigator and the sequence monitors the motion of the heart itself, not rejecting any imaging data at all [63]. 2. Time-resolved (dynamic) 3-dimensional gadolinium imaging (Fig.  1.24) is similar to the static version mentioned above; however, multiple 3-dimensional data sets are obtained in an extremely short period of time (subsecond). These 3-dimensional data sets are acquired continuously and the bolus of gadolinium is followed through the cardiovascular system. This can be performed successfully with parallel imaging (multiple coils used), strong gradients and slew rates, and advanced software and sequences. Each phase of the acquisition can be made into a high-resolution (1 mm or less isotropic voxels) 3-dimensional image. This approach can be used to image physiology (view shunt flow or small connections) as well as determining lung perfusion (e.g., regions of lung with higher flows will be much brighter than regions of lung with low flow or no flow such as with pulmonary embolism). In either technique, it is recommended, as best as possible, to acquire isotropic voxels (voxels with the same dimensions in all three directions). When acquiring the data in this manner, the 3-dimensional image can be manipulated and resliced in any plane and the resulting image would appear as if it was acquired in that orientation. In neonates and infants where smaller spatial resolution is needed, isotropic voxels are also very important but signal to noise takes on a very important role; as long as the voxel sizes are small, the field of view can remain high (i.e., by keeping the matrix high) and signal to noise will be preserved. The advent of a new iron-based contrast agent, ferumoxytol, had changed the game in pediatrics. It has a long half-life of approximately 14 h which means that high-resolution imaging can be done in a longer period of time without loss of signal (as opposed to a first pass technique like cardiac catheterization angiography or ultrasound contrast agents). In a large-scale multicenter safety trial, it has been found safe in children as well as adults [21]. Using the free breathing 3D inversion recovery gradient echo sequence, resolutions of 0.8 mm can be easily reached in less than 5–7  min in young patients, visualizing coronaries in neonates and creating 4D cine images [20]. With the ability to maintain signal, both respiratory and cardiac phases can be imaged, yielding 4D cine imaging at different phases of the respiratory cycle without the need for a navigator for the diaphragm [18].

1  General Principles of Cardiac Magnetic Resonance Imaging

1.8.5.6 Viability (Delayed Enhancement) (Figs. 1.7 and 1.10) [64] It is generally believed that infarcted myocardium is less of an issue in congenital heart disease than it is in adults and there is some truth to this—but just some. Many lesions in pediatric cardiology, either in the native state or postoperatively, are at risk for myocardial necrosis. The list of diseases that may manifest discrete myocardial scarring can be divided into three large categories and examples given below. 1. Native coronary lesions such as anomalous left coronary artery from the pulmonary artery or from the opposite sinus or coronary fistula or sinusoids in lesions such as pulmonary atresia with intact ventricular septum. 2. Surgical coronary manipulation/postsurgical lesions such as transposition of the great arteries after arterial switch operation or after a Ross operation where coronaries are manipulated or patients with single ventricle or repaired tetralogy of Fallot. 3. Acquired pediatric heart disease such as Kawasaki’s disease, myocarditis, or dilated cardiomyopathy. 4. Native myocardial lesions such as hypertrophic cardiomyopathy, Duchenne’s muscular dystrophy, and in tumor characterization (e.g., fibroma). CMR viability imaging, also known as delayed enhancement, has shown to be effective in the detection of myocardial scarring and can be applied to the infant through adult with congenital heart disease [65–67]. With viability imaging, intravenous gadolinium chelate, which can freely distribute in extracellular water but cannot cross intact cellular membranes, is delivered to the myocardium and accumulates in areas of fibrosis due to increased volume of distribution and slower washout kinetics [68]. This imaging technique distinguishes areas of myocardial scarring with high signal intensity in comparison to viable myocardium. As examples, many studies have used delayed enhancement to correlate myocardial scarring with heart function and clinical outcome in different diseases. Babu-Narayan et al. correlated delayed enhancement in transposition of the great arteries patients after surgical repair with age, length of time after surgery, higher right ventricle end-systolic volume index, and lower RV ejection fraction [69]. Babu-Narayan et al. also showed correlation of scarring with increased QRS duration, QT dispersion, and JT dispersion from ECG exams as well as a significantly higher occurrence of arrhythmia/ syncope. Myocardial scarring has also been investigated in tetralogy of Fallot patients after repair [70, 71]. RV delayed enhancement was shown to correlate with decreased exercise tolerance, increased RV indexed end-systolic volume, decreased RV ejection fraction, and more documented clinical arrhythmia. Left ventricular delayed enhancement correlated with more arrhythmia, shorter exercise duration,

31

increased LV indexed end-diastolic and end-systolic volume, and decreased LV ejection fraction. Regions of myocardial delayed enhancement have also been known to occur in patients with hypertrophic cardiomyopathy [72, 73] and the extent of delayed enhancement has been associated with clinical markers of sudden death risk and progression to heart failure [74]. Regions of irreversible myocardial injury will exhibit high signal intensity on T1-weighted images when administered gadolinium which significantly shortens the longitudinal relaxation time. Although the mechanism by which this occurs is open to debate, it is thought that ruptured cell membranes of myocytes allow the gadolinium to be avidly taken up by the scarred myocardium into the intracellular regions; this results in increased tissue concentration of the contrast agent and, hence, an increased signal intensity. In addition, with scar tissue, there is increased interstitial space between collagen fibers, allowing for gadolinium to inculcate itself in these regions and become more concentrated. This is opposed to normally perfused myocardium where the gadolinium is subsequently “washed” out by coronary blood flow. The signal intensity-time curves separate, with the infarcted myocardium gadolinium curve becoming signal intense much before perfused myocardium and remaining highly signal intense after 5–20 min whereas normal myocardium becomes much less so. CMR pulse sequences, first described in the literature in the mid-1980s, have taken advantage of this high concentration and slow washout of contrast agent when attempting to image infarcted myocardium. With segmented inversion recovery fast gradient echo sequences and other techniques such as steady-state free precession, signal intensity differences between normal and infarcted myocardium can be as high as 500%. The technique has been shown to accurately delineate the presence, extent, and location of acute and chronic myocardial infarction. After preliminary scout images and cine sequences are obtained, 0.1–0.2  mmol/kg of intravenous gadolinium is administered. The myocardium is then imaged approximately 5–20  min after this injection; as neonates and children wash out the contrast agent quicker, they are usually imaged on the “sooner side” or if imaged later, have a longer inversion time (TI) (see below). The sequence makes use of a nonselective 180° inversion pulse which spoils all the spins in the myocardium (black on the image) and gives it T1 weighting. The magnetization of tissue goes from +1 to −1 by this process. As both the myocardial and scar tissue begin to recover their spins (enabling the myocardial and scar tissue to “give off” signal), because scar tissue shortens the T1, it recovers signal much quicker than normal myocardium. A time delay is placed after the 180° inversion pulse (TI) to image the ventricle at just the point where the normal myocardium is about to regain signal again (and because the scar tissue recovers spins much quicker, can give off signal). That

M. A. Fogel

32

is to say that the TI is chosen between the nonselective 180° pulse and the center of k-space of the sequence so that the magnetization of normal myocardium is near the zero line (i.e., normal myocardium is dark on the image). This allows for maximizing the difference in signal intensity between scared and normal myocardium; the ventricle is imaged in mid-late diastole. TI time can be manually changed as the acquisition time continues. Collections of lines of k-space generally occur every other heartbeat. There are a number of recent advances that have further refined the ability of CMR delayed enhancement to detect myocardial fibrosis. Some manufacturers have implemented 3-dimensional volumes slabs to allow for thinner slices. Steady-state free precession imaging can be used as opposed to gradient echo imaging to decrease the time needed for image acquisition. In addition, “single-shot” techniques are available (as opposed to segmented versions) to obtain an entire slice in one heartbeat, further allowing for increased coverage of the myocardium in one breath-hold. Motion-­ corrected viability imaging with high signal intensity as well as wideband viability imaging allows for improved fidelity of the images as well as compensating for motion and metallic artifact respectively. Further advances have relied on refining the correct TI time which, as can be surmised, is a critical component to the whole procedure; as mentioned, it is chosen to optimally “null” the myocardium (i.e., the time at which normal myocardium crosses the “zero” point of signal intensity) where the difference in signal intensity between normal and infarcted tissue is maximized. With a TI time that is too short, the normal myocardium will be below the zero point which will cause two issues (a) differences between signal intensities of infracted and normal myocardium will not be maximized and (b) as the image intensity is a function of the magnitude of the magnetization vector, normal myocardium may become hyperenhanced and scar tissue may become nulled if the TI is shortened enough. On the other hand, if the TI time is too long, the normal myocardium will be shades of gray with the scar tissue having high signal as well; as one can see, the contrast between the two tissues will be reduced. Finally, as mentioned already, as gadolinium concentration decreases from the myocardium as time progresses, the TI will have to be adjusted upwards the longer the time after injection. To make better choices of TI times, two advances have improved the process: (1) “TI scouts” have been developed which obtain images at various increments of TI. The imager can choose the TI based on this scout as to image which appears optimal. (2) A phase-sensitive inversion recovery approach can be used which provides consistent contrast between normal and scarred tissue over a wide range of TIs. This “auto viability” technique maintains signal polarity; the inversion recovery preparation pulse and phase reference acquisition are interleaved requiring 2 heartbeats.

There are a few pitfalls with viability imaging. In patients who cannot hold their breath or with arrhythmias, using certain viability sequences, image quality can be degraded although with single-shot viability imaging and motion-­ corrected versions, this is less of an issue. In addition, ghosting artifacts can occur from tissue which have long T1 values, such as pericardial effusion.

1.8.5.7 Multiparametric Mapping (Fig. 1.8) One of the strengths of CMR is the ability to characterize tissue such as discrete fibrosis or tumor characterization. T1, T2, and T2 star mapping has emerged as extensions of this capability which allows for spatially assessing the relaxation times of each one of these spin properties to assess for diffuse fibrosis (extracellular volume), myocardial edema, or iron deposition respectively. A disproportionate accumulation of collagen in the heart is an important factor in the etiologies of heart failure, diastolic dysfunction, and sudden cardiac death or as a result of valve lesions such as aortic stenosis and this generalized increase in collagen can be measured by the T1 relaxation time. Similarly, assessment of myocardial edema in regions near infarcts or in inflammatory states such as myocarditis would be useful to identify and this can be measured by T2 mapping. By the analysis of T2* in patients who may have iron overload (thalassemia, sickle cell disease, etc.), the amount of myocardial iron present can be measured. There are multiple papers which describe these techniques in great detail which is beyond the scope of this chapter and the reader is referred to them [75–77]. The general concept behind, for example, T1 mapping, is acquiring many images with different T1 weightings, and the signal intensities in each are fit to the equation for T1 relaxation. To be more specific, the magnetization of the tissues/spins are either inverted or nulled with a radiofrequency pulse, and the T1-weighted images acquired at different T1 times or times after the inversion or saturation. By doing this over multiple heartbeats and allowing for the tissue to relax, in between a series of tissue excitations, a T1 relaxation curve can be created. When this is performed prior to gadolinium administration (native T1 relaxation time) and after gadolinium (post contrast T1 relaxation), with a blood hematocrit, the extracellular volume may be calculated as a percent by: 1 1 − postcontrast T1 myo native T1 myo ECV = (1 − hematocrit ) 1 1 − postcontrast T1 blood native T1 blood An estimate of the hematocrit from the images, also termed “synthetic hematocrit,” has been touted to be a good substitute for a blood drawn hematocrit [78] but this is con-

1  General Principles of Cardiac Magnetic Resonance Imaging

troversial [79]. Similar techniques are used for T2 and T2 star but without the use of gadolinium. In the CMR protocol, T1 mapping is performed prior to gadolinium administration and then approximately 15 min afterwards or generally a few minutes after viability imaging. T1 mapping has found utility in multiple areas of pediatric congenital and acquired heart disease. For example, in tetralogy of Fallot, expansion of the extracellular space in both the right and left ventricles is present [80, 81] and expansion of left ventricular extracellular space in adult tetralogy of Fallot patients has been linked to adverse events [82]. Similarly, T2 mapping for edema has found utility in the adult world in acute ischemia and in both the pediatric and adult world as one of the criteria for myocarditis [83]. Generally, values over 55–60  ms at 1.5  T are considered abnormal.

33

techniques have been used in the past to overcome these motion-related problems and a history of how modern coronary imaging is performed by CMR is beyond the scope of this chapter; suffice it to note that it has been a long road to the present-day high-quality, high-resolution imaging of the coronary arteries. To compensate for coronary motion, high-quality ECG or vector gating is required; peripheral pulse gating would not be adequate. In addition, if arrhythmias were present and “arrhythmia rejection” algorithms not used, image degradation will be present. In either case, cardiac motion is compensated for by imaging during the “quiescent phase” of the cardiac cycle which is generally mid to late diastole or if the patient has a high heart rate (generally >100 beats/min), end-­ systole. End-systole is advantageous as well in that this is the phase where the coronaries are most filled with blood. To determine the “quiescent phase” of the cardiac cycle, it has T2* for Myocardial Iron Assessment  In brief, the sequence been our practice to perform a 4-chamber view and left venobtains multiple images of the same short axis slice at vari- tricular outflow tract high temporal resolution cine (30–60 ous echo times utilizing a gradient echo sequence. With lon- phases depending upon the heart rate), focusing on the atrioger TEs, the myocardium and liver become increasingly ventricular groove and aortic annular motion respectively. dark. Because iron is ferromagnetic, the magnetic properties When both of these regions are motionless should be considof the myocardium and liver change with increasing iron ered the beginning of the “quiescent phase.” In addition, the concentration, decreasing the measured T2* (which makes length of the “quiescent phase” is measured in both views as the myocardium even darker) relative to normal myocar- well and is timed only as both structures remain motionless; dium. Values 20 mmHg PAWP >15 mmHg PVR >2 WU

Clinical groupsa 1, 3, 4 and 5

2 and 5

2 and 5

Adapted from [3] and [4] mPAP mean pulmonary arterial pressure, PAWP pulmonary arterial wedge pressure, PVR pulmonary vascular resistance, WU wood units a group 1, pulmonary arterial hypertension (PAH); group 2, PH associated with left heart disease; group 3, PH associated with lung diseases and/or hypoxia; group 4, PH associated with pulmonary artery obstructions; group 5, PH with unclear and/or multifactorial mechanisms

pulmonary shunts; Eisenmenger syndrome; repaired defects; and PAH with small or coincidental defects [5]. Whilst pulmonary vascular disease is encountered in patients with Fontan circulation, elevated mPAP is most frequently post-­capillary in origin. Furthermore, PH due to CHD can arise in other WHO groups. For example, group 2 PH can result from congenital mitral stenosis or cor triatriatum and group 4 PH due to congenital pulmonary artery stenosis. Segmental PH, whereby abnormally elevated pressures are found only in discrete areas of the pulmonary vasculature perfused by aorto-pulmonary collaterals, can be found in cases of complex CHD such as pulmonary or tricuspid atresia [6].

8.3 Epidemiology The overall prevalence of PH is estimated at approximately 1% of the global population and increases with age up to 10% of individuals aged over 65 years [7]. Groups 2 and 3 are the commonest and second commonest clinical PH subtypes with 1-year mortalities of 41% and 46%, respectively [1]. Pulmonary arterial hypertension (group 1 PH) is a relatively rare disorder with an estimated prevalence of 10 to 52 cases per million [8]. It should be noted, however, that these estimates of incidence and prevalence of PH are based upon the previous long-standing haemodynamic definition of resting mPAP ≥25 mmHg [9]. Registry data of patients with PAH suggest the proportion of PAH due to CHD ranges from 10% to 20% [8]. The global prevalence of PAH-CHD has previously been estimated to be about 25 people per million in the entire adult population [7]. However, a contemporary nationwide cross-sectional registry study estimated the prevalence of PAH in patients with CHD at 3.2%, resulting in an estimated prevalence of PAH-­

CHD of approximately 100 per million in the general population [10]. Patients with PAH-CHD have a more favourable prognosis than those with idiopathic PAH [11]. Nevertheless, patients with PAH-CHD are more symptomatic and have at least double the risk of mortality than CHD patients without PAH [12, 13].

8.4 Symptoms and Signs The clinical presentation of PH relates to underlying RV dysfunction and, in the initial stages, tends to be related to exertion. Dyspnoea is the commonest symptom, with fatigue and exercise intolerance, pre-syncope, syncope, and angina also being manifest. Given the non-specific nature of symptoms coupled with the relative rarity of PH, late diagnosis is not uncommon with patients having previously consulted a number of healthcare professionals prior to a final diagnosis being established. Chest pain may manifest with mechanisms including RV ischaemia [14], left main stem compression [15], or co-morbid coronary artery disease. Ankle oedema and abdominal distension due to ascites reflect decompensated right heart failure. More rarely, mechanical complications of severe pulmonary arterial dilatation include hoarse voice resulting from left recurrent laryngeal nerve compression and wheeze secondary to large airways compression. Haemoptysis can result from rupture of engorged hypertrophied bronchial arteries that form systemic to pulmonary collateral vessels. Clinical examination findings of raised RV afterload include a left parasternal heave and a loud pulmonary component of the second heart sound. Elevated jugular venous pressure (JVP) reflects right atrial pressure. Murmurs of tricuspid regurgitation and pulmonary regurgitation may be present. Peripheral oedema, ascites, and hepatomegaly are suggestive of RV decompensation. Signs of underlying PH aetiology should also be sought, for example: clinical stigmata of systemic sclerosis including telangiectasia, sclerodactyly, digital ulceration, calcinosis, and Raynaud’s phenomenon; fine inspiratory crackles could suggest the presence of interstitial lung disease; features of chronic liver disease such as jaundice, spider naevi and palmar erythema; and finger clubbing, which has a differential diagnosis that includes cyanotic congenital heart disease, pulmonary venoocclusive disease, interstitial lung disease, chronic lung infections and liver cirrhosis.

8.5 Treatment Strategies Targeted PAH therapies work by lowering pulmonary pressures and target three separate pathways. The nitric oxide pathway may be targeted by phosphodiesterase type 5 inhibi-

139

8  Pulmonary Hypertension

tors (PDE-5i, such as sildenafil and tadalafil) or soluble guanylate cyclase (sGC) stimulators (such as riociguat). The endothelin pathway is targeted by endothelin receptor antagonists (ERA) that include bosentan, ambrisentan and macitentan. The prostacyclin pathway can be targeted with oral selective prostacyclin IP receptor agonists (selexipag) or prostacyclin analogues (epoprostenol, iloprost and ­treprostinil) which have modes of administration including intravenous, inhaled and subcutaneous. Epoprostenol is the only treatment shown to reduce mortality in idiopathic PAH in a single randomized controlled trial [16]. These separate pathways are synergistic and should be concomitantly targeted in a manner analogous to the treatment strategies in systemic hypertension and heart failure. This approach is supported by the AMBITION trial, which showed that initial combination therapy with tadalafil and ambrisentan resulted in a significantly lower risk of clinical-failure events than monotherapy in treatment-naïve PAH patients [17]. Calcium channel blockers are reserved for the small proportion of idiopathic PAH patients with a favourable response to acute vasodilator testing. The successful management of PAH-CHD requires a multidisciplinary team input in specialist centres. Patients with Eisenmenger syndrome and likely also other PAHCHD benefit from advanced PAH therapies [18]. A comprehensive review of targeted therapies in PAH-CHD is beyond the scope of this text and is summarized in relevant published guidelines [5]. There is insufficient evidence to support the use of PAH therapies in groups 2 or 3 PH, with treatment in these groups aimed at the underlying cause. Furthermore, little evidence exists for targeted PH therapy in group 5 PH. Suitable patients with chronic thromboembolic PH and appropriate distribution and burden of disease may benefit from interventions such as pulmonary endarterectomy or balloon pulmonary angioplasty.

Table 8.2  CMR imaging protocol in patients with PH Core protocol (excluding localizer and scout images) Single-plane cine images  – Four-chamber  – RV long-axis  – LV long-axis Volumetric assessment: with a minimum temporal resolution of 40 ms  – Short-axis (with 7–10 mm spacing between slices, depending upon adult or paediatric patients)  – Transaxial (with 5–8 mm spacing between slices, depending upon adult or paediatric patients) Phase-contrast MR imaging  – Ascending aorta  – Main and branch pulmonary arteries Additional protocol (depending upon clinical indication(s) and incident versus follow-up studies)  – MR angiography and perfusion  – Myocardial tissue characterization (especially in adults and in patients with connective tissue disease-associated PH and/or previous acquired left-sided heart disease)

8.6.1 Ventricular Function and Remodelling

8.6.1.1 Volumetric Assessment Cardiovascular magnetic resonance is the reference standard technique for quantification of cardiac chamber size and function in terms of both accuracy and reproducibility [19]. An abundance of studies have shown that CMRderived metrics of RV function and biventricular cavity size are prognostic in both adults and children with PH [20–22]. Additionally, a meta-analysis of CMR studies has shown that volumetric indices of RV size and function predict clinical worsening as well as mortality, an important composite endpoint in PAH clinical trials [23]. Conventionally, the left ventricular (LV) stack of cine images on CMR is acquired in the LV short-­axis imaging plane. Conversely, a transaxial stack of cine images offers improved reproducibility for RV volumetric analysis [24, 25]. This is likely due to the removal of through-­plane longitudinal motion (which con8.6 Role of CMR in PH fers difficulty in delineating RV cavity volumes especially in the basal short-axis cine slices) and easier identification The key aspects of PH management include making the of the tricuspid and pulmonary valves. For these reasons, a diagnosis, establishing the underlying cause, risk stratifica- transaxial stack of cine images is recommended for dedition for the evaluation of disease severity, monitoring disease cated RV volumetric analysis [26]. progression and assessing response to treatment. The well-­ The importance of CMR-derived volumetric parameters established roles of CMR in each of these facets of PH man- in both risk stratification and monitoring the response to agement will be reviewed in this chapter. This will be divided targeted PAH treatment has become increasingly recoginto relevant components of a CMR study in PH (Table 8.2), nized. Indexed RV end-systolic volume (ESVi) adjusted for starting with the most important aspect of CMR assessment age and sex, for example, improves risk stratification for of ventricular function and remodelling followed by CMR 1-year mortality when used in conjunction with either the measures of RV afterload and vascular physiology. The role Registry to EValuate Early And Long-term PAH disease of CMR in shunt assessment will be reviewed and, finally, management (REVEAL) 2.0 risk score calculator or a modnovel CMR techniques that provide new non-invasive ified French Pulmonary Hypertension Registry (FPHR) insights into PH. approach [27]. Additionally, risk assessment of patients

140

with idiopathic PAH at 1-year follow-up based upon CMRderived metrics has been shown to be at least equal to that based upon invasive RHC assessment [28]. Changes in CMR-derived RV ejection fraction (RVEF) with targeted PAH therapy but not PVR have been shown to be associated with survival in patients with PH [29]. Furthermore, deterioration of RV function can still occur after targeted PAH therapy despite reduction in PVR. A change in stroke volume (SV) of 10 mL measured on CMR has been demonstrated to be clinically relevant in patients with PH [30]. This is advantageous as a non-invasive tool for serial patient follow-up and has subsequently been utilized as a primary endpoint in the Right vEntricular remodelling in Pulmonary ArterIal hypeRtension (REPAIR) study [31]. This multicentre trial evaluated the effects of macitentan, an endothelin receptor antagonist, on RV and haemodynamic outcomes in patients with PAH. This growing body of evidence supporting the role of CMR in the management of patients with PAH has culminated in the inclusion of CMR-derived RVEF, indexed SV and RV ESVi in the comprehensive PAH risk assessment tool in the 2022 European Society of Cardiology (ESC) and European Respiratory Society (ERS) PH guidelines [3].

8.6.1.2 Deformation Imaging Strain (the fractional change in the length of a myocardial segment) and strain rate (the rate of deformation over time) are markers of biventricular myocardial deformation and

D. Knight and V. Muthurangu

enable assessment of wall motion in three directions, namely, longitudinal, circumferential and radial. Feature tracking applied to conventionally acquired cine images is a commercially available approach to strain analysis by CMR, analogous to speckle-tracking software applied to echocardiography images and with similar findings (Fig.  8.1). Impairment in strain can detect ventricular dysfunction before changes in global ejection fraction, with evidence of abnormal resting and exercise RV longitudinal strain despite normal RVEF in patients with PAH [32]. Peak global circumferential RV strain rate has also been shown to be impaired in patients with PH and preserved RVEF, with RV global longitudinal strain, global longitudinal strain rate and global circumferential strain rate all independently predictive of outcome [33]. Beyond RV assessment, abnormalities in LV myocardial mechanics in patients with PH and preserved LVEF have also been shown by feature tracking and tissue phase mapping techniques [34–36]. These changes could reflect the importance of ventricular interdependence as well as LV cardiomyocyte atrophy and contractile dysfunction as potential pathophysiological mechanisms of LV dysfunction in PH [37, 38]. Deformation analysis also allows a deeper insight into right atrial function, which is itself not conventionally measured clinically. Indeed, right atrial (RA) dysfunction as evidenced by impaired right atrial strain is associated with decompensated haemodynamics and RV function in PH [39, 40].

Fig. 8.1  Right ventricular strain analysis by CMR feature tracking in a patient with PH. (Reproduced from [80])

8  Pulmonary Hypertension

141

Fig. 8.3  Abnormal interventricular septal dynamics on a short-axis mid-ventricular cine image in a patient with PH. Note the septum bows Fig. 8.2  Prominent inferior RV insertion point and mid-septal mid-­ towards the left ventricle at end-systole, indicative of raised RV wall late gadolinium enhancement in a patient with PH afterload

8.6.1.3 Myocardial Tissue Characterization Late gadolinium enhancement (LGE) is well described in PH at the RV insertion points and the mid-wall of the interventricular septum (Fig.  8.2). Autopsy studies have suggested that these findings may represent exaggerated myocardial disarray and plexiform fibrosis, possibly as a consequence of RV hypertrophy and abnormal interventricular septal dynamics [41]. However, whilst RV insertional LGE is suggestive of the presence of PH and has been shown to predict clinical worsening in this condition, it does not provide additive prognostic information to CMR-derived RVEF [42, 43]. It is also a non-specific finding that is present in other pathologies such as hypertrophic heart muscle disease and so is not diagnostic for PH as an isolated finding. Non-contrast (native) myocardial T1 mapping and extracellular volume (ECV) quantification have also been studied in PH, albeit their clinical utility is less wellestablished [44–50]. Perhaps unsurprisingly and akin to the description of insertional LGE in PH, native myocardial T1 is particularly elevated at the RV insertion points and is associated with measures of abnormal interventricular septal dynamics [46, 48]. Whilst RV insertion point T1 correlates with severity of PH, it has not shown prognostic capacity in PH and also does not discriminate between patients with or without PH [44, 48]. Overall, multiparametric myocardial mapping is more challenging in the thin-walled RV.  Nevertheless, RV ECV has been described as significantly higher in patients with PH versus patients with LV systolic dysfunction without PH [45]. Furthermore, whilst RV ECV was related to RV dilatation and dysfunction, it was independently associated with clinical diagnosis and thus may describe changes in the RV myocardium beyond conventional CMR-derived functional metrics.

8.6.2 Measures of Afterload and Vascular Physiology 8.6.2.1 Interventricular Septal Configuration The interventricular septal dynamics are highly revealing of loading conditions of the RV (Fig. 8.3). Many studies have shown significant correlations between CMR measures of interventricular septal curvature and invasively derived mPAP and PVR in both adults and children [51–56]. Furthermore, measures of septal curvature have been shown to track changes in pulmonary haemodynamics during vasodilator testing, providing a potential non-invasive method for assessing response to therapy in PH [54]. Sophisticated CMR analysis has also enabled a more detailed evaluation of the mechanistic basis behind these deleterious interventricular septal dynamics and their impact on cardiac function. Deformation analysis using tagged CMR cine images shows a ‘post-systolic’ RV contractile phase of the RV [57]. This causes relatively higher RV versus LV pressures with consequent leftward bowing and deviation of the interventricular septum during LV isovolumic relaxation period [53, 58]. This discordant pattern of biventricular contraction and relaxation gives rise to ventricular interdependence, manifesting as impaired LV filling in early diastole and ultimately reducing LV stoke volume. This may be a reason underlying why CMR indices of LV early diastolic function are so important with respect to functional capacity and clinical worsening in patients with PH [36]. 8.6.2.2 Pulmonary Arterial Geometry The pulmonary arterial tree is characterized by its low pressure and low resistance as well as a high degree of distensibility. The relative area change of the pulmonary artery (PA) represents its elasticity, is inversely proportional to stiffness

142

and is simple to measure at CMR, reflecting the advantages of CMR as a non-invasive, dynamic, cross-sectional imaging modality. Relative area change is calculated as the percentage increase in PA area during the cardiac cycle. There is a curvilinear relationship between mPAP and PA relative area change, reflecting hysteresis of vessel biomechanics. Irrespective of this non-linear relationship, the stiffness of the proximal pulmonary arterial tree is associated with more severe PH and is an independent predictor of outcome in PAH [21, 59, 60]. It should also be noted that changes in stiffness as PH progresses are associated with vessel remodelling. However, there are currently no specific reference ranges or diagnostic clinical thresholds for relative area change in PH.

8.6.2.3 Magnetic Resonance Angiography and Perfusion The demonstration of pulmonary arterial thromboembolic changes or mismatched pulmonary perfusion defects coupled with elevated pulmonary arterial pressures is required in order to diagnose chronic thromboembolic PH (CTEPH). Conventionally, non-invasive imaging of the pulmonary vasculature is performed using computed tomography pulmonary angiograms (CTPA) and nuclear ventilation/perfusion (V/Q) scans. Alternatively, anatomical assessment of the proximal pulmonary arterial tree can be provided by magnetic resonance angiography (MRA), and MRI can also be used to visualize pulmonary parenchymal perfusion (Fig.  8.4). The advantages of MRI include the absence of ionizing radiation that is encountered with CTPA and V/Q scans along with providing an alternative modality for

D. Knight and V. Muthurangu

patients who cannot receive the iodinated contrast agents required for CTPA.  However, a particular limitation to be considered of both MRA and MRI-based perfusion imaging is the long breath-hold times which are particularly challenging for breathless patients. The V/Q scan still remains the preferred imaging test for CTEPH screening, with a normal study effectively excluding the condition [61]. Nevertheless, retrospective studies of three-dimensional dynamic contrast-enhanced lung perfusion MRI have shown promise for diagnosing CTEPH, with a similar sensitivity when compared with planar V/Q scans and a higher sensitivity when compared with single photon emission computed tomography (SPECT) scans [62, 63]. The results of ongoing prospective studies of dynamic contrast-­ enhanced MRI-derived pulmonary perfusion to fully assess their diagnostic performance in clinical practice are awaited to determine the place of this technique in the diagnostic pathway of CTEPH [64]. Contrast-enhanced MRA has shown promise in imaging the proximal pulmonary arteries, but CTPA remains the non-invasive test of choice [65, 66].

8.6.3 Quantitative Shunt Assessment 8.6.3.1 The Role of Phase-Contrast MR The quantification of blood flow velocity and volume is an important part of the routine assessment of PH by CMR. Cardiac output is reduced in PH, and abnormal flow characteristics are seen in pulmonary vessels due to the altered mechanical properties of the pulmonary vascular bed. Great vessel flow is quantified by velocity-encoded phase-­ contrast MRI (PCMR), can be acquired during free-­breathing and has been extensively validated. However, swirling rather than laminar or plug flow patterns occurs in dilated pulmonary arteries which can result in inaccuracies in through-­ plane flow quantification. Therefore, proximal branch PA flows as well as main PA flow should be acquired in PH studies to provide corroborative data. Great vessel flow assessment by PCMR allows the calculation of cardiac output, shunt flow (Qp:Qs), right and left lung perfusion ratios, collateral flow and aortic and pulmonary valvular regurgitation. Furthermore, biventricular stroke volume assessment by cine imaging combined with great vessel flow evaluation by PCMR permits the quantification of atrioventricular valvular regurgitation.

8.6.4 Novel CMR Techniques to Assess PH Fig. 8.4  Magnetic resonance pulmonary perfusion in a patient with chronic thromboembolic PH. Note the lack of perfusion in the majority of the left lung

8.6.4.1 4D Flow and Vortices Three-dimensional spatial encoding combined with three-­ directional velocity-encoded PCMR, termed 4D flow, offers

8  Pulmonary Hypertension

143

Fig. 8.5  Velocity colour-encoded three-dimensional vector representation of vortical blood flow along the main pulmonary artery in a patient with known PH (images from left to right correspond to systole through to diastole). (Reproduced from [69])

the abilities to measure and visualize the temporal evolution of complex blood flow patterns, namely, vortices, within an acquired 3D volume. Abnormal flow characteristics are demonstrated in the right heart and pulmonary circulation in patients with PH, with a change in flow from laminar or plug to a helical flow pattern (Fig. 8.5). The duration of abnormal vortical flow in the main PA correlates with mPAP and has been demonstrated as a non-invasive estimate of PA pressure [67–70].

8.6.4.2 Wave Intensity Analysis Abnormal PA wave reflections contribute to increased RV afterload, the most obvious in clinical practice being the mid-systolic notching of the PA flow velocity waveform. Abnormal wave reflections can be quantified using wave intensity analysis (WIA) [71, 72]. This allows the measurement of the magnitude, timing, direction (forward or backward) and type (compression or expansion waves) of vascular waves. Traditionally, WIA has been limited by the requirement for invasively derived pressure waveforms with simultaneous measurement of flow. Instead, high temporal resolution PCMR-derived flow and area waveforms can be used to non-invasively quantify travelling waves in the pulmonary arteries. For example, CMR-based WIA has demonstrated that the size of the backwards compression wave is increased in PH and discriminates between CTEPH subtypes [73]. Furthermore, CMR-derived WIA indices have been shown to be predictive of functional worsening in children with PAH [74]. 8.6.4.3 MR-Augmented Right Heart Catheterization Whilst a plethora of CMR metrics can be used to indicate the presence of PH or provide surrogates of pulmonary haemodynamics, pressure can only be directly obtained by invasive RHC. However, there are limitations of the measurement of CO by RHC by both thermodilution and Fick methods, such

as the presence of shunts and valvular insufficiency. Alternatively, phase-contrast magnetic resonance (PCMR) provides the reference standard for flow analysis which can be combined with concurrent invasive pressure data [75]. Moreover, CMR has the advantages of freedom from radiation, a pertinent point in children and patients who require repeated procedures, and soft-tissue visualization which is of particular utility in complex anatomy. Therefore, attention has grown in RHC under MRI guidance (MR-RHC). Combined MR-RHC is not a novel concept but has been relatively slow to clinically adopt, largely due to reliance on expensive infrastructure and relatively long procedure times [75, 76]. More recently, MR-RHC has been performed safely in conventional CMR environments with clinically acceptable procedure times including in patients with PH [77–79]. Combined MR-RHC also has the convenient benefit of acquiring important additional data such as cardiac chamber size and function quantification in a single comprehensive procedure.

8.7 Conclusion Numerous studies confirm the clinical importance and utility of CMR in patients with PH.  It is the reference standard imaging modality for the assessment of biventricular volumes and function as well as for the non-invasive quantification of blood flow. This is particularly valuable in PH given that the response of the RV to elevated afterload is of greater prognostic significance than the change in afterload itself. Nevertheless, the clinical evaluation of PH still requires the direct measurement of pulmonary haemodynamics for diagnosis and the assessment of response to PH-targeted therapies. There are many CMR measures that provide estimates of PA pressure but, ultimately, cannot replace the requirement for RHC.  However, combining CMR with invasive pressure measurement by RHC has emerged as a clinically

144

feasible technique. This permits the accurate quantification of pulmonary vascular resistance, provides a comprehensive assessment of PH and affords potential for more novel assessments of afterload such as wave intensity analysis and impedance spectra. It is likely that the role of CMR will continue to grow and evolve in the field of PH, including the increasing use of CMR-derived biomarkers as endpoints in clinical trials of PH therapies.

References 1. Mathai SC, Ryan JJ. The growing burden of pulmonary hypertension in the modern era: a zebra no more? Circ Cardiovasc Qual Outcomes. 2018;11(2):e004536. 2. Chin KM, Kim NH, Rubin LJ.  The right ventricle in pulmonary hypertension. Coron Artery Dis. 2005;16(1):13–8. 3. Humbert M, Kovacs G, Hoeper MM, Badagliacca R, Berger RMF, Brida M, et  al. 2022 ESC/ERS guidelines for the diagnosis and treatment of pulmonary hypertension. Eur Heart J. 2022;43(38):3618–731. 4. Simonneau G, Montani D, Celermajer DS, Denton CP, Gatzoulis MA, Krowka M, et  al. Haemodynamic definitions and updated clinical classification of pulmonary hypertension. Eur Respir J. 2019;53(1):1801913. 5. Baumgartner H, De Backer J, Babu-Narayan SV, Budts W, Chessa M, Diller GP, et  al. 2020 ESC guidelines for the management of adult congenital heart disease. Eur Heart J. 2021;42(6):563–645. 6. Dimopoulos K, Diller GP, Opotowsky AR, D’Alto M, Gu H, Giannakoulas G, et  al. Definition and management of segmental pulmonary hypertension. J Am Heart Assoc. 2018;7(14):e008587. 7. Hoeper MM, Humbert M, Souza R, Idrees M, Kawut SM, Sliwa-­ Hahnle K, et al. A global view of pulmonary hypertension. Lancet Respir Med. 2016;4(4):306–22. 8. Hoeper MM, Simon RGJ.  The changing landscape of pulmonary arterial hypertension and implications for patient care. Eur Respir Rev. 2014;23(134):450–7. 9. Hoeper MM, Bogaard HJ, Condliffe R, Frantz R, Khanna D, Kurzyna M, et al. Definitions and diagnosis of pulmonary hypertension. J Am Coll Cardiol. 2013;62(25 Suppl):D42–50. 10. van Riel AC, Schuuring MJ, van Hessen ID, Zwinderman AH, Cozijnsen L, Reichert CL, et  al. Contemporary prevalence of pulmonary arterial hypertension in adult congenital heart disease following the updated clinical classification. Int J Cardiol. 2014;174(2):299–305. 11. Benza RL, Miller DP, Barst RJ, Badesch DB, Frost AE, McGoon MD. An evaluation of long-term survival from time of diagnosis in pulmonary arterial hypertension from the REVEAL registry. Chest. 2012;142(2):448–56. 12. Lowe BS, Therrien J, Ionescu-Ittu R, Pilote L, Martucci G, Marelli AJ.  Diagnosis of pulmonary hypertension in the congenital heart disease adult population impact on outcomes. J Am Coll Cardiol. 2011;58(5):538–46. 13. Verheugt CL, Uiterwaal CS, van der Velde ET, Meijboom FJ, Pieper PG, van Dijk AP, et al. Mortality in adult congenital heart disease. Eur Heart J. 2010;31(10):1220–9. 14. Gómez A, Bialostozky D, Zajarias A, Santos E, Palomar A, Martínez ML, et  al. Right ventricular ischemia in patients with primary pulmonary hypertension. J Am Coll Cardiol. 2001;38(4):1137–42. 15. Galiè N, Saia F, Palazzini M, Manes A, Russo V, Bacchi Reggiani ML, et al. Left main coronary artery compression in patients with

D. Knight and V. Muthurangu pulmonary arterial hypertension and angina. J Am Coll Cardiol. 2017;69(23):2808–17. 16. Barst RJ, Rubin LJ, Long WA, McGoon MD, Rich S, Badesch DB, et al. A comparison of continuous intravenous epoprostenol (prostacyclin) with conventional therapy for primary pulmonary hypertension. N Engl J Med. 1996;334(5):296–301. 17. Galiè N, Barberà JA, Frost AE, Ghofrani HA, Hoeper MM, McLaughlin VV, et al. Initial use of ambrisentan plus tadalafil in pulmonary arterial hypertension. N Engl J Med. 2015;373(9):834–44. 18. Dimopoulos K, Inuzuka R, Goletto S, Giannakoulas G, Swan L, Wort SJ, et al. Improved survival among patients with eisenmenger syndrome receiving advanced therapy for pulmonary arterial hypertension. Circulation. 2010;121(1):20–5. 19. Schulz-Menger J, Bluemke DA, Bremerich J, Flamm SD, Fogel MA, Friedrich MG, et  al. Standardized image interpretation and post-processing in cardiovascular magnetic resonance −2020 update: Society for Cardiovascular Magnetic Resonance (SCMR): board of trustees task force on standardized post-processing. J Cardiovasc Magn Reson. 2020;22(1):19. 20. Moledina S, Pandya B, Bartsota M, Mortensen KH, McMillan M, Quyam S, et  al. Prognostic significance of cardiac magnetic resonance imaging in children with pulmonary hypertension. Circ Cardiovasc Imaging. 2013;6(3):407–14. 21. Swift AJ, Capener D, Johns C, Hamilton N, Rothman A, Elliot C, et al. Magnetic resonance imaging in the prognostic evaluation of patients with pulmonary arterial hypertension. Am J Respir Crit Care Med. 2017;196(2):228–39. 22. van Wolferen SA, Marcus JT, Boonstra A, Marques KM, Bronzwaer JG, Spreeuwenberg MD, et al. Prognostic value of right ventricular mass, volume, and function in idiopathic pulmonary arterial hypertension. Eur Heart J. 2007;28(10):1250–7. 23. Alabed S, Shahin Y, Garg P, Alandejani F, Johns CS, Lewis RA, et  al. Cardiac-MRI predicts clinical worsening and mortality in pulmonary arterial hypertension: a systematic review and meta-­ analysis. JACC Cardiovasc Imaging. 2021;14(5):931–42. 24. Alfakih K, Plein S, Bloomer T, Jones T, Ridgway J, Sivananthan M.  Comparison of right ventricular volume measurements between axial and short axis orientation using steady-state free precession magnetic resonance imaging. J Magn Reson Imaging. 2003;18(1):25–32. 25. Atalay MK, Chang KJ, Grand DJ, Haji-Momenian S, Machan JT, Sheehan FH. The transaxial orientation is superior to both the short axis and horizontal long axis orientations for determining right ventricular volume and ejection fraction using Simpson’s method with cardiac magnetic resonance. ISRN Cardiol. 2013;2013:268697. 26. Kramer CM, Barkhausen J, Flamm SD, Kim RJ, Nagel E. Standardized cardiovascular magnetic resonance (CMR) protocols 2013 update. J Cardiovasc Magn Reson. 2013;15:91. 27. Lewis RA, Johns CS, Cogliano M, Capener D, Tubman E, Elliot CA, et  al. Identification of cardiac magnetic resonance imaging thresholds for risk stratification in pulmonary arterial hypertension. Am J Respir Crit Care Med. 2020;201(4):458–68. 28. van der Bruggen CE, Handoko ML, Bogaard HJ, Marcus JT, Oosterveer FPT, Meijboom LJ, et  al. The value of hemodynamic measurements or cardiac MRI in the follow-up of patients with idiopathic pulmonary arterial hypertension. Chest. 2021;159(4):1575–85. 29. van de Veerdonk MC, Kind T, Marcus JT, Mauritz GJ, Heymans MW, Bogaard HJ, et al. Progressive right ventricular dysfunction in patients with pulmonary arterial hypertension responding to therapy. J Am Coll Cardiol. 2011;58(24):2511–9. 30. van Wolferen SA, van de Veerdonk MC, Mauritz GJ, Jacobs W, Marcus JT, Marques KM, et al. Clinically significant change in stroke volume in pulmonary hypertension. Chest. 2011;139(5):1003–9. 31. Vonk Noordegraaf A, Channick R, Cottreel E, Kiely DG, Marcus JT, Martin N, et al. The REPAIR study: effects of macitentan on RV

8  Pulmonary Hypertension structure and function in pulmonary arterial hypertension. JACC Cardiovasc Imaging. 2021;15(2):240–53. 32. Lin ACW, Seale H, Hamilton-Craig C, Morris NR, Strugnell W.  Quantification of biventricular strain and assessment of ventriculo-­ventricular interaction in pulmonary arterial hypertension using exercise cardiac magnetic resonance imaging and myocardial feature tracking. J Magn Reson Imaging. 2019;49(5):1427–36. 33. de Siqueira ME, Pozo E, Fernandes VR, Sengupta PP, Modesto K, Gupta SS, et al. Characterization and clinical significance of right ventricular mechanics in pulmonary hypertension evaluated with cardiovascular magnetic resonance feature tracking. J Cardiovasc Magn Reson. 2016;18(1):39. 34. Homsi R, Luetkens JA, Skowasch D, Pizarro C, Sprinkart AM, Gieseke J, et al. Left ventricular myocardial fibrosis, atrophy, and impaired contractility in patients with pulmonary arterial hypertension and a preserved left ventricular function: a cardiac magnetic resonance study. J Thorac Imaging. 2017;32(1):36–42. 35. Kallianos K, Brooks GC, Mukai K, Seguro de Carvalho F, Liu J, Naeger DM, et  al. Cardiac magnetic resonance evaluation of left ventricular myocardial strain in pulmonary hypertension. Acad Radiol. 2018;25(1):129–35. 36. Knight DS, Steeden JA, Moledina S, Jones A, Coghlan JG, Muthurangu V.  Left ventricular diastolic dysfunction in pulmonary hypertension predicts functional capacity and clinical worsening: a tissue phase mapping study. J Cardiovasc Magn Reson. 2015;17:116. 37. Gan C, Lankhaar JW, Marcus JT, Westerhof N, Marques KM, Bronzwaer JG, et al. Impaired left ventricular filling due to right-to-­ left ventricular interaction in patients with pulmonary arterial hypertension. Am J Physiol Heart Circ Physiol. 2006;290(4):H1528–33. 38. Manders E, Bogaard HJ, Handoko ML, van de Veerdonk MC, Keogh A, Westerhof N, et al. Contractile dysfunction of left ventricular cardiomyocytes in patients with pulmonary arterial hypertension. J Am Coll Cardiol. 2014;64(1):28–37. 39. Leng S, Dong Y, Wu Y, Zhao X, Ruan W, Zhang G, et al. Impaired cardiovascular magnetic resonance-derived rapid semiautomated right atrial longitudinal strain is associated with decompensated Hemodynamics in pulmonary arterial hypertension. Circ Cardiovasc Imaging. 2019;12(5):e008582. 40. Tello K, Dalmer A, Vanderpool R, Ghofrani HA, Naeije R, Roller F, et al. Right ventricular function correlates of right atrial strain in pulmonary hypertension: a combined cardiac magnetic resonance and conductance catheter study. Am J Physiol Heart Circ Physiol. 2020;318(1):H156–h64. 41. Bradlow WM, Assomull R, Kilner PJ, Gibbs JS, Sheppard MN, Mohiaddin RH. Understanding late gadolinium enhancement in pulmonary hypertension. Circ Cardiovasc Imaging. 2010;3(4):501–3. 42. Freed BH, Gomberg-Maitland M, Chandra S, Mor-Avi V, Rich S, Archer SL, et al. Late gadolinium enhancement cardiovascular magnetic resonance predicts clinical worsening in patients with pulmonary hypertension. J Cardiovasc Magn Reson. 2012;14(1):11. 43. Swift AJ, Rajaram S, Capener D, Elliot C, Condliffe R, Wild JM, et al. LGE patterns in pulmonary hypertension do not impact overall mortality. JACC Cardiovasc Imaging. 2014;7(12):1209–17. 44. Chen YY, Yun H, Jin H, Kong H, Long YL, Fu CX, et  al. Association of native T1 times with biventricular function and hemodynamics in precapillary pulmonary hypertension. Int J Cardiovasc Imaging. 2017;33(8):1179–89. 45. Mehta BB, Auger DA, Gonzalez JA, Workman V, Chen X, Chow K, et al. Detection of elevated right ventricular extracellular volume in pulmonary hypertension using accelerated and navigator-gated look-locker imaging for cardiac T1 estimation (ANGIE) cardiovascular magnetic resonance. J Cardiovasc Magn Reson. 2015;17:110. 46. Reiter U, Reiter G, Kovacs G, Adelsmayr G, Greiser A, Olschewski H, et al. Native myocardial T1 mapping in pulmonary hypertension:

145 correlations with cardiac function and hemodynamics. Eur Radiol. 2017;27(1):157–66. 47. Roller FC, Wiedenroth C, Breithecker A, Liebetrau C, Mayer E, Schneider C, et  al. Native T1 mapping and extracellular volume fraction measurement for assessment of right ventricular insertion point and septal fibrosis in chronic thromboembolic pulmonary hypertension. Eur Radiol. 2017;27(5):1980–91. 48. Saunders LC, Johns CS, Stewart NJ, Oram CJE, Capener DA, Puntmann VO, et al. Diagnostic and prognostic significance of cardiovascular magnetic resonance native myocardial T1 mapping in patients with pulmonary hypertension. J Cardiovasc Magn Reson. 2018;20(1):78. 49. Spruijt OA, Vissers L, Bogaard HJ, Hofman MB, Vonk-Noordegraaf A, Marcus JT.  Increased native T1-values at the interventricular insertion regions in precapillary pulmonary hypertension. Int J Cardiovasc Imaging. 2016;32(3):451–9. 50. Wang J, Zhao H, Wang Y, Herrmann HC, Witschey WRT, Han Y.  Native T1 and T2 mapping by cardiovascular magnetic resonance imaging in pressure overloaded left and right heart diseases. J Thorac Dis. 2018;10(5):2968–75. 51. Bouchard A, Higgins CB, Byrd BF 3rd, Amparo EG, Osaki L, Axelrod R.  Magnetic resonance imaging in pulmonary arterial hypertension. Am J Cardiol. 1985;56(15):938–42. 52. Dellegrottaglie S, Sanz J, Poon M, Viles-Gonzalez JF, Sulica R, Goyenechea M, et al. Pulmonary hypertension: accuracy of detection with left ventricular septal-to-free wall curvature ratio measured at cardiac MR. Radiology. 2007;243(1):63–9. 53. Marcus JT, Gan CT, Zwanenburg JJ, Boonstra A, Allaart CP, Gotte MJ, et  al. Interventricular mechanical asynchrony in pulmonary arterial hypertension: left-to-right delay in peak shortening is related to right ventricular overload and left ventricular underfilling. J Am Coll Cardiol. 2008;51(7):750–7. 54. Pandya B, Quail MA, Steeden JA, McKee A, Odille F, Taylor AM, et al. Real-time magnetic resonance assessment of septal curvature accurately tracks acute hemodynamic changes in pediatric pulmonary hypertension. Circ Cardiovasc Imaging. 2014;7(4):706–13. 55. Roeleveld RJ, Marcus JT, Faes TJ, Gan TJ, Boonstra A, Postmus PE, et  al. Interventricular septal configuration at mr imaging and pulmonary arterial pressure in pulmonary hypertension. Radiology. 2005;234(3):710–7. 56. Swift AJ, Rajaram S, Hurdman J, Hill C, Davies C, Sproson TW, et  al. Noninvasive estimation of PA pressure, flow, and resistance with CMR imaging: derivation and prospective validation study from the ASPIRE registry. JACC Cardiovasc Imaging. 2013;6(10):1036–47. 57. Mauritz GJ, Marcus JT, Westerhof N, Postmus PE, Vonk-­ Noordegraaf A.  Prolonged right ventricular post-systolic isovolumic period in pulmonary arterial hypertension is not a reflection of diastolic dysfunction. Heart. 2011;97(6):473–8. 58. Marcus JT, Vonk Noordegraaf A, Roeleveld RJ, Postmus PE, Heethaar RM, Van Rossum AC, et  al. Impaired left ventricular filling due to right ventricular pressure overload in primary pulmonary hypertension: noninvasive monitoring using MRI.  Chest. 2001;119(6):1761–5. 59. Gan CT, Lankhaar JW, Westerhof N, Marcus JT, Becker A, Twisk JW, et  al. Noninvasively assessed pulmonary artery ­ stiffness predicts mortality in pulmonary arterial hypertension. Chest. 2007;132(6):1906–12. 60. Ray JC, Burger C, Mergo P, Safford R, Blackshear J, Austin C, et al. Pulmonary arterial stiffness assessed by cardiovascular magnetic resonance imaging is a predictor of mild pulmonary arterial hypertension. Int J Cardiovasc Imaging. 2019;35(10):1881–92. 61. Kim NH, Delcroix M, Jais X, Madani MM, Matsubara H, Mayer E, et al. Chronic thromboembolic pulmonary hypertension. Eur Respir J. 2019;53(1):1801915.

146 62. Johns CS, Swift AJ, Rajaram S, Hughes PJC, Capener DJ, Kiely DG, et al. Lung perfusion: MRI vs. SPECT for screening in suspected chronic thromboembolic pulmonary hypertension. J Magn Reson Imaging. 2017;46(6):1693–7. 63. Rajaram S, Swift AJ, Telfer A, Hurdman J, Marshall H, Lorenz E, et al. 3D contrast-enhanced lung perfusion MRI is an effective screening tool for chronic thromboembolic pulmonary hypertension: results from the ASPIRE registry. Thorax. 2013;68(7):677–8. 64. Lasch F, Karch A, Koch A, Derlin T, Voskrebenzev A, Alsady TM, et al. Comparison of MRI and VQ-SPECT as a screening test for patients with suspected CTEPH: CHANGE-MRI study design and rationale. Front Cardiovasc Med. 2020;7:51. 65. Ley S, Kauczor HU, Heussel CP, Kramm T, Mayer E, Thelen M, et al. Value of contrast-enhanced MR angiography and helical CT angiography in chronic thromboembolic pulmonary hypertension. Eur Radiol. 2003;13(10):2365–71. 66. Ley S, Ley-Zaporozhan J, Pitton MB, Schneider J, Wirth GM, Mayer E, et al. Diagnostic performance of state-of-the-art imaging techniques for morphological assessment of vascular abnormalities in patients with chronic thromboembolic pulmonary hypertension (CTEPH). Eur Radiol. 2012;22(3):607–16. 67. Reiter G, Reiter U, Kovacs G, Kainz B, Schmidt K, Maier R, et al. Magnetic resonance-derived 3-dimensional blood flow patterns in the main pulmonary artery as a marker of pulmonary hypertension and a measure of elevated mean pulmonary arterial pressure. Circ Cardiovasc Imaging. 2008;1(1):23–30. 68. Reiter G, Reiter U, Kovacs G, Olschewski H, Fuchsjäger M. Blood flow vortices along the main pulmonary artery measured with MR imaging for diagnosis of pulmonary hypertension. Radiology. 2015;275(1):71–9. 69. Reiter U, Kovacs G, Reiter C, Kräuter C, Nizhnikava V, Fuchsjäger M, et al. MR 4D flow-based mean pulmonary arterial pressure tracking in pulmonary hypertension. Eur Radiol. 2021;31(4):1883–93. 70. Schäfer M, Barker AJ, Kheyfets V, Stenmark KR, Crapo J, Yeager ME, et  al. Helicity and vorticity of pulmonary arterial flow in patients with pulmonary hypertension: quantitative analysis of flow formations. J Am Heart Assoc. 2017;6(12):e007010. 71. Lau EM, Abelson D, Dwyer N, Yu Y, Ng MK, Celermajer DS.  Assessment of ventriculo-arterial interaction in pulmonary

D. Knight and V. Muthurangu arterial hypertension using wave intensity analysis. Eur Respir J. 2014;43(6):1804–7. 72. Su J, Manisty C, Parker KH, Simonsen U, Nielsen-Kudsk JE, Mellemkjaer S, et al. Wave intensity analysis provides novel insights into pulmonary arterial hypertension and chronic thromboembolic pulmonary hypertension. J Am Heart Assoc. 2017;6(11):e006679. 73. Quail MA, Knight DS, Steeden JA, Taelman L, Moledina S, Taylor AM, et  al. Noninvasive pulmonary artery wave intensity analysis in pulmonary hypertension. Am J Physiol Heart Circ Physiol. 2015;308(12):H1603–11. 74. Schäfer M, Wilson N, Ivy DD, Ing R, Abman S, Browne LP, et al. Noninvasive wave intensity analysis predicts functional worsening in children with pulmonary arterial hypertension. Am J Physiol Heart Circ Physiol. 2018;315(4):H968–h77. 75. Razavi R, Hill DL, Keevil SF, Miquel ME, Muthurangu V, Hegde S, et al. Cardiac catheterisation guided by MRI in children and adults with congenital heart disease. Lancet. 2003;362(9399):1877–82. 76. Kuehne T, Yilmaz S, Schulze-Neick I, Wellnhofer E, Ewert P, Nagel E, et al. Magnetic resonance imaging guided catheterisation for assessment of pulmonary vascular resistance: in vivo validation and clinical application in patients with pulmonary hypertension. Heart. 2005;91(8):1064–9. 77. Knight DS, Kotecha T, Martinez-Naharro A, Brown JT, Bertelli M, Fontana M, et al. Cardiovascular magnetic resonance-guided right heart catheterization in a conventional CMR environment—predictors of procedure success and duration in pulmonary artery hypertension. J Cardiovasc Magn Reson. 2019;21(1):57. 78. Rogers T, Ratnayaka K, Khan JM, Stine A, Schenke WH, Grant LP, et al. CMR fluoroscopy right heart catheterization for cardiac output and pulmonary vascular resistance: results in 102 patients. J Cardiovasc Magn Reson. 2017;19(1):54. 79. Ratnayaka K, Kanter JP, Faranesh AZ, Grant EK, Olivieri LJ, Cross RR, et al. Radiation-free CMR diagnostic heart catheterization in children. J Cardiovasc Magn Reson. 2017;19(1):65. 80. Alabed S, Garg P, Johns CS, Alandejani F, Shahin Y, Dwivedi K, et  al. Cardiac magnetic resonance in pulmonary hypertension-an update. Curr Cardiovasc Imaging Rep. 2020;13(12):30.

9

Tetralogy of Fallot Michael A. Quail, Vivek Muthurangu, and Andrew M. Taylor

9.1 Introduction Tetralogy of Fallot (TOF) is the most common form of cyanotic congenital heart disease, occurring in 1  in 3600 live births [1]. Complete repair of TOF was devised over 50 years ago (first reported by Lillehei in 1954) and can result in complete intra-cardiac repair in early infancy [2]. There are excellent short- and medium-term survival rates, and increasingly the 25-year actuarial survival for patients repaired before their fifth birthday is greater 90% of the expected survival rate [3] though the annualized risk of death triples in the third postoperative decade [4]. Late morbidity and mortality, in particular related to pulmonary incompetence, have been observed in many patients long after total repair. Initially, pulmonary incompetence was believed to be a relatively benign condition, with few problems associated with right ventricular volume loading. However, it has become clear that chronic pulmonary incompetence and right ventricular (RV) volume loading can cause RV dysfunction, which can in turn lead to symptoms of reduced exercise tolerance, increased risk of atrial and ventricular tachyarrhythmia, and sudden death. This has led to an Supplementary Information The online version contains supplementary material available at https://doi.org/10.1007/978-­3-­031-­29235-­4_9.

increasing proportion of patients requiring operative replacement of incompetent and/or stenosed pulmonary valves and conduits. The accurate quantification of pulmonary incompetence and stenosis and their effects on the right and left ventricles and great vessel anatomy/pathology is therefore crucial: Data that can be provided by cardiovascular magnetic resonance (CMR) imaging. In this chapter, we will provide an overview of TOF, its treatment and its assessment with CMR.  Ultimately, CMR may enable improvements in the management of late complications through appropriate surveillance and treatment and may extend the survival and quality of life of patients.

9.2 Morphology Morphologically, the principal defect of this condition is antero-cephalad deviation of the muscular outlet septum resulting in the tetrad of ventricular septal defect (VSD), pulmonary outflow tract obstruction, overriding aorta, and RV hypertrophy (Fig. 9.1). In TOF, the malaligned outlet septum serves to narrow the sub-pulmonary outflow tract and simultaneously create an interventricular defect that is overridden by the aortic valve apparatus.

M. A. Quail Cardiovascular Imaging and Adult Congenital Heart Disease, Great Ormond Street Hospital for Children, London, UK e-mail: [email protected] V. Muthurangu Cardiovascular Imaging and Physics, UCL Institute of Cardiovascular Science, University College London, London, UK e-mail: [email protected] A. M. Taylor (*) Cardiovascular Imaging, UCL Institute of Cardiovascular Science and Great Ormond Street Hospital for Children, London, UK e-mail: [email protected]

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. A. Syed, R. H. Mohiaddin (eds.), Magnetic Resonance Imaging of Congenital Heart Disease, https://doi.org/10.1007/978-3-031-29235-4_9

147

148

M. A. Quail et al.

Overriding aorta

Pulmonary stenosis

Patch repairing septal defect Ventricular septal defect Transannular patch to right ventricular outflow tract

Infundibular stenosis

Right ventricular hypertrophy

Fig. 9.1  Schematic diagram of unrepaired TOF (copyright belongs to Gemma Price)

The VSD in TOF is usually nonrestrictive and subaortic; however, it may occasionally extend into the sub-pulmonary region. Its inferior and anterior borders are comprised of the limbs of the septomarginal trabeculation, while the superior border is formed from the deviated outlet septum and posteriorly the anteroseptal leaflet of the tricuspid valve. Obstruction of the right ventricular outflow tract (RVOT) in TOF is most frequently in the form of infundibular stenosis (45%) with obstruction rarely isolated to the pulmonary valve alone (10%), and more usually a combination of the two (15%). In its most severe form, the pulmonary valve is atretic (15%) [5]. The deviated outlet septum combined with hypertrophied septoparietal trabeculations contributes significantly to subvalvular obstruction and inevitably requires division at surgery. The pulmonary valve is frequently stenotic with thickened and tethered leaflets requiring valvotomy or placement of a transannular patch if the annulus diameter is deemed inadequate (Fig. 9.2). However, the resultant sequelae of pulmonary incompetence are now recognized to be a significant problem later in life. Stenosis of the origin of the left and/or right pulmonary arteries (PAs) is a common finding and may require resection if localized or placement of a separate patch is more diffusely hypoplastic. In a proportion of patients, there may be pulmonary atresia, particularly at an infundibular or valvular level and associated major aortopulmonary collateral arteries (MAPCAs). The preferred approach in this situation may be the placement of a right ventricle to pulmonary artery (RV-PA) conduit.

Fig. 9.2  Schematic diagram of repaired TOF (copyright belongs to Gemma Price)

The individual components of the tetrad clearly show marked variability, especially the nature of pulmonary stenosis and the extent of the VSD; these variations account for the spectrum of clinical severity, and their surgical treatment can significantly influence later outcomes.

9.2.1 Associated Anomalies Significant associated cardiac defects are uncommon. The most frequent associated lesions include right-sided aortic arch, atrial septal defect, patent ductus arteriosus, atrioventricular septal defects, and additional VSDs. Less commonly, there may be a persistent left-sided superior vena cava, anomalous origin of the left anterior descending coronary artery, or aortopulmonary collaterals if severe pulmonary stenosis. Associated syndromic conditions in which TOF may occur as a major manifestation include DiGeorge syndrome (22q11.2 deletion) and Alagille syndrome (JAG1 mutation); however, the condition more often occurs as an isolated defect.

9.3 Clinical Presentation The clinical presentation of TOF in infancy varies depending on the degree of RVOT obstruction. Typically, the infant will present with cyanosis due to right-to-left shunt, and diagnosis is then established by echocardiography. Surgery is usu-

9  Tetralogy of Fallot

ally performed around 3–6  months as cyanosis progresses, often without a prior palliative shunt procedure. In the infant with TOF and pulmonary atresia, severe cyanosis is seen immediately after birth. The complications observed in patients beyond the immediate postoperative period relate to surgical residua and progression of the “unnatural” history of the disease. Common problems requiring re-intervention include severe ­ pulmonary regurgitation, residual outflow tract obstruction, and conduit failure. Less commonly a residual VSD, particularly at the posteroinferior margin of the patch, and severe tricuspid regurgitation require reoperation. Pulmonary regurgitation, in contrast to other residua, is remarkably well tolerated. However, over time, it produces its deleterious effects through volume overload of the right ventricle. It is associated with pathological RV dilatation and dysfunction, decreased exercise tolerance, sudden cardiac death, and ventricular arrhythmia associated with prolonged QRS duration. Pulmonary valve replacement (PVR) can ameliorate the volume overload imposed by pulmonary regurgitation. However, the optimal timing for this procedure remains unclear, as homograft prostheses have a limited life span and further reoperations may be required should these fail.

9.3.1 Primary Surgical Repair Early primary repair is preferred because it shortens the period the patient is exposed to hypoxemia, right ventricular (RV) pressure overload, and subsequent RV hypertrophy. It is performed with a low perioperative mortality, with reported midterm survival in the modern era of 97% at 7 years of postoperative follow-up [3]. The surgical approach has also shifted from a transventricular approach to a trans-atrial/trans-pulmonary approach, with the aim of preserving RV structure and myocardial function and reducing the potential side effects of an RV ventriculotomy (coronary artery damage, RV contractile reduction, scar arrhythmia).

149

9.3.2 Late Surgical Treatment: Pulmonary Valve Replacement Surgical pulmonary valve replacement (SPVR) can be performed using a variety of valve materials including cadaveric homografts and man-made conduits with excellent outcomes [6]. Additional procedures may also be necessary and can be performed during the same cardiopulmonary bypass including resection of residual trabeculations, extensive enlargement of pulmonary arteries, closure of a patent foramen ovale, residual atrial and/or ventricular septal defects, tricuspid valve annuloplasty, and aortic valve replacement. PVR is now commonly performed with a degree of RVOT refashioning [7] to reduce RVOT aneurismal dilatation and optimally align the RVOT and new pulmonary trunk.

9.3.3 Percutaneous Pulmonary Valve Replacement In patients who develop stenosis of their surgical PVR (homograft or conduit stenosis), percutaneous transcatheter stenting is possible. This can prolong conduit life and postpone reoperation. However bare-metal stenting can potentially convert the pressure-overloaded ventricle to one of volume overload through relief of obstruction and introduction of free pulmonary incompetence. In 2000, percutaneous pulmonary valve replacement (PPVR) was described, whereby a new pulmonary valve was placed into a dysfunctional RV-PA prosthetic conduit [8]. Over the following decade, this technology has been accepted into clinical practice, with over many thousands of devices now implanted throughout the world [9–11]. Multiple devices can be used in this setting including the Melody™ (Medtronic Inc., Minneapolis, MN, USA) (Fig.  9.3) and Edwards SAPIEN transcatheter heart valve [12]. This method offers a minimally invasive alternative to open-heart surgery for RVOT/pulmonary trunk dysfunction in children and adults by restoring acceptable RV loading conditions, in particular when compared to bare-metal stenting of pulmonary conduit stenosis [13].

150

a

M. A. Quail et al.

b

c

Fig. 9.3  PPVI device (a). Device viewed in water bath showing the valve open during forward flow (b) and closed during reverse flow (c) note tri-leaflet morphology of the valve

It must be noted that one complication of the Melody device is infective endocarditis (IE) that occurs more frequently than in patients undergoing SPVR—survival free of IE by Kaplan–Meier at 5 years of 98.7% for homografts vs. 84.9% for Melody [14].

9.4 CMR Imaging As patients with congenital heart disease in general, and TOF in particular, are increasingly surviving longer, appropriate follow-up and surveillance for complications are becoming more important. In this regard, echocardiography has been the mainstay of investigation—providing important information regarding intracardiac anatomy, assessing valvular competence, and ventricular function. However, the technique is limited by postoperative restriction of acoustic windows, inability to provide sufficient hemodynamic infor-

mation, and because extracardiac anatomy, such as branch PAs, can be difficult to assess. CMR has become an important modality in the assessment of TOF because the technique can provide three-­ dimensional (3D) anatomy of the right-sided cardiac and vascular structures crucial to adequate clinical assessment. CMR can provide reliable serial hemodynamic information noninvasively and, unlike echocardiography, can quantify pulmonary regurgitation volume and flow in branch PAs. CMR can also provide myocardial tissue fingerprinting, using focal late-gadolinium enhancement to assess areas of myocardial scarring, and T1 mapping and extracellular volume (ECV) imaging to assess diffuse fibrosis. Ultimately, CMR provides a large field of view and unlimited choice of imaging planes and is much less operator dependent than echocardiography. A current limitation of CMR is its inability to measure hemodynamic pressures.

9  Tetralogy of Fallot

9.4.1 Indications CMR is primarily used for the assessment of patients with TOF late after primary repair for the assessment of significant residual lesions that are identified at routine follow-up. Patients also undergo CMR assessment routinely before transfer from pediatric to adult congenital cardiac services to provide baseline anatomic and functional information. The exact frequency of serial scans remains to be determined; repeat scans are usually performed early if significant lesions are present, particularly where surgical or catheter intervention may be required.

151

zation. This results in two significant limitations: image blurring due to cardiac motion reduces the ability of this technique to visualize intracardiac anatomy, and the size of individual vessels represents an average size over the cardiac cycle which can lead to underestimation of systolic dimensions. A further disadvantage of gadolinium-enhanced MR angiography is that fast-moving turbulent blood causes signal dropout, leading to overestimation of stenoses.

9.4.2.2 Cine Imaging Cine imaging using balanced-SSFP, or fast gradient echo sequences, gives multiphase data that shows myocardial or valve motion over the entire cardiac cycle. These cines may have up to 40 frames per cardiac cycle, a temporal resolution 9.4.2 CMR Sequences adequate for accurate physiological representation. The balanced-­SSFP sequences are now used as standard, as they 9.4.2.1 3D Imaging provide improved blood pool homogeneity throughout the The 3D capabilities of CMR play a key role for pediatric and cardiac cycle. Balanced-SSFP images have significantly adult patients with congenital heart disease. There are two higher contrast-to-noise ratios allowing better detection of conventional methods of acquiring 3D data. One uses angio- the endocardial border than traditional gradient echo graphic techniques with gadolinium-based contrast agents sequences [18]. The technique lends itself particularly to that can be injected via any peripheral vein. The other uses a assessment of TOF allowing qualitative assessment of car3D balanced-SSFP sequence, which is respiratory and car- diac chambers, pulmonary valvular function, and vascular diac gated, but does not require contrast [15]. Both data sets anatomy. are acquired in such a way to give isotropic voxels, so that Cine images are acquired in equal-width slices perpenthe images can be viewed with the same spatial resolution in dicular to the long axis of the heart, from base to apex, in any anatomical plane. These data can be used during the scan order to assess cardiac function and measure the ventricular to plan image planes for further scanning as well as during volumes (Fig.  9.4). The post-processing of cine images to the reporting phase to assess 3D relationships between struc- calculate ventricular volumes and function is performed off-­ tures, quantify vessel size, and view morphology. The high-­ line, using commercially available or open-source software. signal, isotropic 3D images that are achieved using The segmentation of the blood pool and myocardial border gadolinium-contrast angiography allow complex modeling can be performed manually, or by using automated signal of structures so that interventional techniques can be opti- thresholding techniques. There is currently a wide range of mized [16]. software available and a wide variation in segmenting pracGadolinium-enhanced MR angiography is particularly tice and procedures. A fundamental issue, particularly for useful in the assessment of complex pulmonary vascular pediatric patients and those with congenital disease, is that of anatomy associated with TOF (Movie 9.1), where the pul- inclusion or noninclusion of the trabeculae in the blood pool. monary vascular bed may be supplied with blood flow from If a simple endocardial contour is drawn and the trabeculae other sources, such as surgically created shunts or aortopul- ignored and included in the blood pool, the manual segmenmonary collateral vessels. Delineation of all sources of pul- tation process is more efficient and more reproducible [19]. monary blood supply and the size and morphology of the However, this leads to erroneously large volume estimates pulmonary arteries are essential as surgical and transcatheter for the ventricles and prohibits internal validation of stroke procedures are often required to augment effective pulmo- volumes using great arterial flow volumes. nary blood flow or to eliminate sources of excessive pulmonary blood flow [17]. Gadolinium contrast-enhanced MR 9.4.2.3 Flow Assessment angiography relies on the T1 shortening effect of dilute gad- Accurate quantification of flow volume is crucial in patients olinium. A 3D data set is acquired at the peak of the gado- with known or suspected congenital heart disease. For vollinium bolus. Sequences are designed to ensure that tissues ume quantification, we favor a free-breathing, velocity without gadolinium enhancement are suppressed, resulting encoded phase-contrast sequence with a temporal resolution in prominence of contrast-enhanced vessels. Images are of at least 30 frames per cardiac cycle. Slice positioning and acquired in a single breath-hold, without cardiac synchroni- velocity encoding must be optimized [20]. If these parame-

152

M. A. Quail et al.

Fig. 9.4  Short-axis, b-SSFP imaging stack through the ventricles of a patient with TOF, starting at the base (left to image) passing from left to right top to bottom toward the apex (right bottom image). This is the

diastolic stack—red contour LV end-diastolic volume (EDV), blue— RV EDV, green LV mass contour

ters are rigorously controlled, flow can be assessed in large and small arteries and systemic and pulmonary veins. Aortic and pulmonary valve regurgitant volumes can be directly measured (Fig.  9.5). Phase contrast flow sequences also enable the profiling of flow acceleration jets, with velocity estimation. More importantly, with appropriate combinations of arterial and venous flow volume assessment, the technique allows accurate assessment of interatrial, interventricular, arterial, and venous shunt volumes. In the context of atrioventricular valve regurgitation, knowledge of the ventricular stroke volume, combined with knowledge of the forward arterial flow volume from that ventricle allows for calculation of mitral or tricuspid valve regurgitant fraction. For every patient in whom ventricular function is quantified, the practice of our unit is to undertake great arterial flow vol-

ume assessment to guide the volumetric analysis. This greatly enhances the accuracy and reproducibility of our reporting procedure.

9.4.2.4 Black-Blood Imaging Black-blood spin-echo pulse sequences are seldom used in the CMR protocol nowadays, however, can still play a role in the assessment of TOF. They are good for the morphology of the blood vessels and cardiac chambers, in particular when turbulent flow at the site of stenosis reduces the accuracy of b-SSFP or MR angiography images. Black-blood imaging is also good at elucidating the relationship between airway and blood vessels. In TOF, this allows accurate assessment of branch PA abnormalities including hypoplasia, stenosis, and non-confluence, which are otherwise missed by echocardiography.

9  Tetralogy of Fallot

a

c

153

b

d

300 Aorta Main PA

Flow (mL/s)

200

100

0

-100

0

200

400

600

Time (ms)

Fig. 9.5  MR phase contrast velocity mapping in pulmonary regurgitation—(a) diastolic, modulus image showing non-coaptation of the valve leaflets, (b) systolic phase contrast image with white forward flow

(note black flow in the descending aorta posteriorly), (c) diastolic phase contrast showing pulmonary regurgitation (black), (d) flow curve over the cardiac cycle

Following treatment of PA stenoses by intravascular stenting, or following PPVR, gradient echo sequences suffer from serious metal artifacts due to T2* field inhomogeneity. Though spin-echo is less susceptible to metal artifacts, this does not necessarily lead to a better assessment of stents and may possibly lead to false reassurance as to their patency [21]. CT imaging may be required to definitively exclude intra-stent stenosis.

9.4.2.5 Tissue Characterization Late-gadolinium enhancement (LGE) demonstrates large focal areas of fibrosis or other myocardial abnormality and is frequently observed in patients with TOF (Fig.  9.6) [22]. This is particularly so in areas that reflect the surgical repair-­ RVOT scar, site of VSD repair, and ventriculotomy site; however, areas of LGE are also seen remote to sites of surgery. Importantly, LGE correlates to RV systolic function

154

M. A. Quail et al.

Fig. 9.6  Example of severe RV LGE extent. The RV is divided into six segments (yellow numbers 1–6). Regions of RV LGE were scored according to linear extent (0 1/4 no enhancement, 1 1/4 up to 2 cm, 2 1/4 up to 3 cm, 3 1/4 3 or more cm in length) and number of trabeculations enhanced including the moderator band (0 1/4 no enhancement, 1

1/4 1 trabeculation, 2 1/4 2–4). From Ghonim S, et  al. JACC Cardiovascular Imaging 2022; Feb;15(2):257–268. doi: 10.1016/j. jcmg.2021.07.026 used under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/ licenses/by/4.0/)

and dilatation and, ultimately, may provide some prognostic information for timing PVR in repaired TOF. Other forms of tissues characterization using either T1 mapping [23] or ECV imaging [24] remain in their infancy for use in routine TOF CMR assessment. Large-scale studies to understand the role of tissue fingerprinting in TOF are required.

ated with adverse outcome in patients with repaired TOF [25] after midterm (8.9 years) follow-up in 100 patients. In normal subjects, RV ejection fraction increases during stress while in patients with TOF RV function remains unchanged or reduced during stress. The responses of subjects with chronic pulmonary incompetence to stress may be able to provide important prognostic indictors to help with the timing of pulmonary valve replacement in this patient ­population. Importantly, the use of real-time imaging to assess flow and function [26] during exercise may mean that such studies can now be carried out, without the need for pharmacological stress or the need to breath-hold during exercise, making exercise MR more acceptable to the patient.

9.4.2.6 Stress CMR Imaging Stress CMR imaging can be performed either with exercise (specifically designed MR bicycle) or dobutamine administration and has been used as means of assessing global RV function in response to increased workload. An abnormal ventricular response to dobutamine stress has been associ-

9  Tetralogy of Fallot

155

Table 9.1  Routine clinical imaging protocol Scout

Sequence Single shot b-SSFP

Ventricular long-axis b-SSFP cine images, breath(RV and LV) held, ECG gated AV valves Four-chamber view Short-axis stack

b-SSFP cine images, breathheld, ECG gated b-SSFP cine images, breathheld, ECG gated b-SSFP cine images, breathheld, ECG gated

MR angiography

Gadolinium injection. Breath held, no ECG gating

LV and RV outflow tracts Bi-ventricular LGE

b-SSFP cine images, breathheld, ECG gated 2D fast low angle shot inversion recovery with myocardial nulling Free breathing, respiratory navigated, ECG-gated. Data acquisition in diastole Through-­plane velocity mapping, ECG gated, non-breath-­held

3D b-SSFP

Great vessel flow

Planning 48 images in 3 orthogonal planes—16 in each plane Orthogonal plane through long axis of ventricle from AV valve to ventricular apex, planned from axial scout images Orthogonal plane parallel to AV valve, planned from axial stack Orthogonal plane across AV valve orifices, planned from AV valves image Orthogonal plane at AV valve hinge points with inclusion of basal blood pool, planned from end-diastolic frame of four-­chamber cine Planned on axial scout images

LV planned from AV valves cine. RV planned from axial stack Short axis stack, 4-chamber and RVOT images copied from appropriate planes Planned on axial scout images for sagittal orientation Planned from outflow tract images (aortic root, pulmonary trunk). AV values and 3D whole heart (branch PAs, SVC, IVC). Measure pulmonary venous flow if aortopulmonary collateral flow needed

Purpose Isocentering To set up subsequent image planes AV valve function, assessment of ventricular volumes Planning four-chamber Atrial/ventricular size and function, AV valve regurgitation Ventricular volume calculation, assessment of septum, and outflow tracts

Angiographic views of large and small vessels. Second pass allows assessment of systemic and pulmonary venous anatomy Outflow tract morphology, subjective assessment of semilunar valve function Focal scar imaging

High-resolution intracardiac anatomy Visualize position of coronary artery origins, and proximal course Vessel flow volume/velocity, calculation of regurgitant fraction and volume. Evaluate shunts

b-SSFP balanced steady-state free precession, RV right ventricle, LV left ventricle, AV atrioventricular, LGE late gadolinium, RVOT right ventricular outflow tract, PAs pulmonary arteries, SVC superior vena cava, IVC inferior vena cava

9.4.3 Clinical Imaging Protocol (Table 9.1)

9.5 CMR Findings in TOF 9.5.1 Pulmonary Valve and RV Assessment The MR assessment of TOF requires comprehensive evaluation of the entire heart; however, emphasis is rightly placed on the evaluation of the RV and pulmonary valve. The RV outflow tract is visualized by aligning a plane that passes through the pulmonary trunk (or conduit) and the RV ­inferiorly from the axial stack. An alternative RVOT plane is a sagittal or sagittal-oblique view through the pulmonary trunk and descending aorta. Pulmonary incompetence can be assessed using a plane perpendicular to the RVOT views described, just above the pulmonary valve.

9.5.1.1 Pulmonary Incompetence Pulmonary incompetence to some degree is a predominant feature of TOF late after repair. CMR can image the regurgi-

tant jet in three dimensions and can quantify the regurgitant volume or describe it as a regurgitant fraction. This quantification is important in clinical decision-making regarding catheter or surgical valve replacement. Turbulence of blood regurgitating through the pulmonary valve in diastole causes de-phasing and signal loss in gradient echo cine imaging (Movie 9.2). This facilitates a qualitative gradation of the regurgitant jet: grade 1, signal loss close to the valve; grade 2, signal loss extending into the proximal chamber; grade 3, signal loss filling the whole of the proximal chamber; and grade 4,  signal loss in the receiving chamber throughout the relevant half of the cardiac cycle. The most accurate assessment of PR however is achieved by through-plane velocity-encoded phase-contrast imaging. Instantaneous flow volumes in the PA are calculated by multiplying the PA contour area (drawn manually) with the spatial average flow within this contour (Fig. 9.5). Total forward and retrograde (regurgitant) flow in a cardiac cycle can be calculated by integrating the instantaneous flow volumes for all frames. This technique has been validated in TOF by comparing it to the differ-

156

ences in left and right stroke volume, which can also be used to quantify pulmonary regurgitation if there is no other valve regurgitation or shunt. The regurgitant fraction is calculated by dividing pulmonary retrograde flow by pulmonary forward flow ×100.

9.5.1.2 Pulmonary Stenosis and RVOT Obstruction Pulmonary stenosis and RVOT obstruction are significant and common residual lesions following TOF repair. Using MR velocity mapping and the modified Bernoulli equation, the gradient across a stenosis can be calculated. This technique is comparable to Doppler echocardiography but is not limited by acoustic windows, allowing measurement of the velocity jet in any plane. Imaging is usually performed using a combination of through-plane (perpendicular to jet) and in-­ plane (parallel to jet) imaging. The latter is used to initially define the jet, with subsequent through-plane images at the site of maximum velocity.

M. A. Quail et al.

a

b

9.5.1.3 Systolic Ventricular Function Systolic ventricular function in TOF is based on CMR ventricular volumetric assessment. The complex geometry of the RV means that no single imaging plane is well suited to RV assessment. We recommend the use of the short axis; however, the interface between right ventricle and right atrium can be difficult to assess, and long-axis imaging may help. 9.5.1.4 Diastolic Ventricular Function Diastolic ventricular function can also be assessed using CMR. Though the late forward flow of so-called restrictive physiology can be seen, RV time-volume curves can be created using either cine ventricular volume data or by combining phase contrast velocity maps of flow through the tricuspid and pulmonary valve. These have not only been used to demonstrate impaired diastolic function in the RV and LV of patients with TOF but also how early diastolic filling can improve following relief of conduit stenosis [27]. Again, further studies of diastolic properties, with both CMR and echocardiography, are required to define the prognostic use of this information.

9.5.2 Unrepaired TOF In the assessment of the patient with unrepaired TOF, the anatomical diagnosis will usually have been made previously by echocardiography. CMR will therefore aim to confirm the characteristic morphology (Fig.  9.7) and assist

Fig. 9.7  Unrepaired TOF in neonate. (a) Histological specimen of showing VSD, overriding aorta (arrowhead), infundibular stenosis (asterisks), and RV hypertrophy (arrow). (b) Comparison CMR image—black-blood coronal oblique view

preoperative planning by defining any additional complex anatomy and physiology that may influence the surgical procedure. Assessment should include: • RV outflow tract obstruction—Is narrowing of the sub-­ pulmonary infundibulum fixed or dynamic? • Pulmonary valve morphology—Is valve sparing surgery possible? • Quantify peak velocity across RV outflow tract. • Identify and measure any branch pulmonary stenosis/ hypoplasia. • Identify and measure any palliative shunts. • Measure differential lung perfusion. • Quantify ventricular volumes, mass, and function.

9  Tetralogy of Fallot

157

a

c

e

b

d

f

Fig. 9.8  Repaired TOF in an adolescent. (a, b) Four-chamber b-SSFP image showing RV dilatation in diastole (dia) and RV hypertrophy in systole (sys,*). (c, d) b-SSFP image showing dilated and aneurysmal RVOT (arrow) following transannular patch repair in oblique sagittal

• Check aortic root for dilation. • Exclude additional VSDs.

9.5.3 Repaired TOF Patients who have undergone surgical repair of TOF will have had repair of VSD and relief of the RVOT obstruction. The latter may have involved surgical resection of infundibular muscle bundles and/or the insertion of a transannular patch if the annular diameter was deemed inadequate. Significant pulmonary regurgitation is almost always encountered following this procedure, though mild/moderate regurgitation is well tolerated. Pulmonary regurgitation may be increased by proximal or distal pulmonary artery stenosis, and chronic severe/free pulmonary regurgitation may lead to RV dilatation and dysfunction. The CMR study aims to comprehensively define the hemodynamic status of the patient (Fig. 9.8).

view (c) diastole and (d) systole—note the paradoxical increase in aneurysm size during systole. (e) 3D MR angiogram reconstruction of RVOT aneurysm (arrow). (f) b-SSFP image through an intact VSD patch (arrow)

The following should be considered: • Describe RV outflow tract and pulmonary trunk anatomy (Fig. 9.9 and Movie 9.2). • Identify any RV outflow tract aneurysm (Fig.  9.8 and Movie 9.2). • Quantify pulmonary valve regurgitation (Fig. 9.5). • Assess proximal and distal pulmonary arteries for stenoses (Fig. 9.9 and Movie 9.1). • Assess biventricular function, volume, and mass (Fig. 9.4 and Movie 9.3). • Assess for LGE in both ventricles and the RVOT (Fig. 9.6). • Check for the presence of MAPCAs. • Measure Qp:Qs—assess for residual shunts, e.g., residual VSD and MAPCAs. • Check aortic root for dilatation. • Assess course of coronary arteries, which may be in proximity PPVR implantation sites, important to know prior to SPVR (Movie 9.4).

158

M. A. Quail et al.

Fig. 9.9  Variations in pulmonary trunk and branch pulmonary artery anatomy. Rapid prototyping models (reconstructed from CMR contrast-­ enhanced angiography data) of 12 patients with TOF assessed 10–15 years after early complete repair

9.6 Clinical Use of CMR The most commonly encountered complication of repaired TOF is severe RV dilatation and dysfunction secondary to free pulmonary incompetence. The dilemma of when to treat patients with free pulmonary incompetence, presenting late after repair of right ventricular outflow tract obstruction, is one that faces all congenital heart disease clinicians. Although there is clear data to suggest that, in the long term, pulmonary incompetence is detrimental, leading to an increased incidence of adverse events (death, sustained arrhythmias, increasing symptoms), the conventional thinking has been that the benefit of treating free pulmonary incompetence is outweighed by the potential risk of surgical pulmonary valve replacement and the lack of longevity of this treatment (conduit dysfunction within 10–15 years that exposes patients to multiple operations over their life) (Fig.  9.10) [28]. PVR has therefore often been left until patients develop symptoms; however, once symptoms develop, there is the potential that symptomatic improvement after surgery may be limited due to the fact that the right ventricle has been chronically exposed to pulmonary incompetence. More recently, there has been a shift in this risk/benefit continuum. Operative mortality and morbidity are now small with modern operative methods and postoperative care [6], and conduit life can now be extended using new non-invasive percutaneous approaches to treat conduit dysfunction.

9.6.1 The Use of CMR to Select Patients with Severe Pulmonary Incompetence for PVR Current data from CMR assessment of patients with severe pulmonary incompetence have demonstrated elevated RV end-diastolic and end-systolic volumes and reduced RV ejection fraction compared to normal. Furthermore, there is increasing evidence that RV function may be irreversibly compromised by such long-term changes. This is exemplified by three findings that have been demonstrated by CMR. Firstly, RV ejection fraction has been shown to be significantly lower in patients with both RV pressure and volume overload as compared with RV pressure overload alone. Secondly, an abnormal RV response to stress has been demonstrated in patients with TOF and pulmonary incompetence. And finally, there appears to be no, or limited, improvement in RV function (ejection fraction at rest) following PVR. Indeed, one of the few studies to demonstrate an improvement in RV ejection fraction following PVR was in patients with moderately dilated right ventricles [29]. Despite this lack of marked improvement in RV function by correcting pulmonary incompetence, PVR does reduce RV dimensions (Fig. 9.11), and if performed before an RV EDV index of 160 mL/m2 or an RV ESV index of 80 mL/m2 RV dimensions can be normalized in the majority of patients [30–32]. These numbers form the basis for both the AHA/ ACC [33] and ESC [34] guidelines for when to consider PVR in patients with repaired TOF (Table 9.2). These num-

9  Tetralogy of Fallot

159

Benefits Regression of RV volumes Improved NYHA class Improved LVEF (particularly if bad at baseline)

Risks Operative risk: 0.9–(4%) Risk of multiple re-interventions (with increasing risk)

Increased risk of endocarditis

No effect

Uncertainties

Peak exercise capacity Right ventricular EF Maternal and foetal risks during pregnancy

Arrhythmia-risk Risk of heart failure Risk of sudden death Timing of re-intervention

Fig. 9.10  Risks, benefits, and uncertainties about pulmonary valve replacement (PVR). LVEF left ventricular ejection fraction, NYHA New  York Heart Association, RV right ventricular. With permission

from Greutmann M.  European Heart Journal 2016; 37:836–839 doi:10.1093/eurheartj/ehv634

bers need to be taken with caution as they represent studies done in patients of varying age groups, with varying original operations and varying PVR surgery (RVOT aneurysmal reduction or not). Furthermore, as there is limited standardization for the measurement of RV volumes, the quantitative measure of RV dimensions can vary between centers. It may be that as the RV dilates, interaction through the septum reduces LV filling [27] and reduces LV function. Importantly, there are suggestions that this may be happening at even moderate RV dilatation, because following PVR, LV EDV increases with a subsequent increase in LV function in some studies [29, 35]. Furthermore, a recent systematic review of outcomes shows that impaired right and left ventricular function is the most consistent independent predictor of disease progression in repaired ToF [36]. Hence, large-scale prospective studies are still necessary to provide important information to guide optimal clinical decision making in this area. Research is required to define CMR prognostic factors for PVR timing, and importantly, other information from echocardiography, exercise testing, assessment of neurohormonal activation, ECG data, and stress CMR may be useful for decision-making [29]. There is increasing evidence from systematic review that LGE may act as marker of progressive impairment of myocardial function with higher amount of right ventricular LGE related to higher right ventricular volumes, lower ejection fraction, and a higher pulmonary regurgitant fraction [37].

Importantly, LGE in TOF patients appears to correlate with the onset of arrhythmias, with a recent [38].

9.6.2 Which Treatment Option? Once a PVR has been clinically justified, there is now a choice between surgery and PPRI for approximately 15–20% of patients with TOF. Technical suitability for PPVR can be defined by CMR using the protocol previously defined, but with focus on three main areas:

9.6.2.1 RVOT/Pulmonary Trunk Size and Distensibility PPVR can be performed in RVOT/pulmonary trunks that range from 14 to 22 mm. At the lower end of the spectrum, conduits need to be of an adequate size to allow for sufficient opening without residual gradients. At the upper end, the device can only be expanded to a maximal diameter of 22 mm—any larger and valve leaflet coaptation may fail, or the device may embolize. This precludes PPVI in dilated anatomies (Fig.  9.9). Conduit sizes can be gleaned from operative reports; however, conduits can become smaller (or larger) over time, and in order to have a full understanding of the anatomy of the outflow tract, CMR with 3D capabilities is crucial. Although the CMR-derived 3D reconstructions can be used to define size, it is important to realize that these

160

M. A. Quail et al.

Fig. 9.11  Response to surgical PVR after 1 year. Four-chamber and short-axis views showing marked RV volume reduction and increased LV volumes

reconstructions are performed on data acquired in diastole, or from non-ECG-gated data and, thus, maximal dimensions of very distensible anatomies may be underestimated. Cine imaging of the RVOT/pulmonary trunk in both long and short axes overcomes this problem, enabling the measurement of the maximum diameter of the site at which PPVI may be attempted. If the results of MRI are doubtful or borderline, balloon sizing of the RVOT can be performed at the time of catheterization.

9.6.2.2 RVOT/Pulmonary Morphology The 3D information from CMR can be used to visualize the best site for device anchorage in the RVOT/pulmonary trunk. Furthermore, certain shapes are not suitable for safe implantation of the device [16]. A morphological classification has been created according to measurements of 3D reconstructions of the RVOT [39]. Importantly, a pyramidal morphology, meaning that the RVOT funnels down toward the pulmonary bifurcation, is not suitable for PPVr because

161

9  Tetralogy of Fallot Table 9.2  Summary of international guidelines for pulmonary valve replacement PVR indication ESC 2020 [34] PVR is recommended in symptomatic patients with severe PR and/or at least moderate RVOTO In patients with no native outflow tract, catheter intervention (PPVR) should be preferred if anatomically feasible PVR should be considered in asymptomatic patients with severe PR and/or RVOTO when one of the following criteria is present:  1. Decrease in objective exercise capacity  2. Progressive RV dilation to RVESVi ≥80 mL/m2, and/or RVEDVi ≥160 mL/m2, and/or progression of TR to at least moderate  3. Progressive RV systolic dysfunction  4. RVOTO with RVSP >80 mmHg ACC 2018 [33] Moderate or greater PR and cardiovascular symptoms not otherwise explained Moderate or greater PR and any two of the following:  1. Mild or moderate RV or LV systolic dysfunction  2. Severe RV enlargement (RVEDV ≥160 mL/ m2, RVESVi ≥80 mL/m2, or RVEDV ≥2 × LVEDV)  3. RVSP due to RVOT obstruction ≥2/3 systemic pa  4. Progressive reduction in objective exercise tolerance Moderate or greater PR and sustained ventricular tachyarrhythmias

Level of Class evidence I

C

I

C

IIa

C

I

B

IIa

B

9.6.2.3 Proximity of the Proximal Coronary Arteries The proximity of the proximal coronary arteries to the RVOT/pulmonary trunk has to be assessed (Movie 9.4). On CMR 3D whole-heart b-SSFP images, the anatomical relationship of the coronary arteries and the proposed ­implantation site can be judged. In addition, aortic root angiography is performed at the time of catheterization. On biplane projection, the relationship between the coronaries and the pulmonary artery can be judged. In case the CMR assessment or aortic root angiography cannot fully rule out the risk for coronary compression, simultaneous high-­ pressure balloon inflation in the implantation site and selective coronary angiography are performed [40]. Importantly, when angiography and simultaneous balloon inflation are performed, it is crucial to expand the RVOT to a therapeutic size. This maneuver is only meaningful when the conduit is expanded with a high-pressure balloon up to the diameter that will be reached post-PPVI.

9.6.3 Acute Hemodynamic Results and Implications for Biventricular Function Following PPVI

IIb

C

PVR pulmonary valve replacement, PR pulmonary regurgitation, RVOTO right ventricular outflow tract obstruction, PPVR percutaneous pulmonary valve replacement, RV right ventricle, RVESVi right ventricular end-systolic volume, indexed for BSA, RVEDVi right ventricular end-diastolic volume, indexed for BSA, LV left ventricle, RVSP right ventricular systolic pressure, pa pressure ESC Class I—Evidence and/or general agreement that a given treatment or procedure is beneficial, useful, effective II—Conflicting evidence and/or a divergence of opinion about usefulness/efficacy of the given treatment or procedure    a—Weight of evidence/opinion is in favor of usefulness/efficacy    b—Usefulness/efficacy is less well established by evidence/ opinion Level of evidence A—Data derived from multiple randomized clinical trial or meta-analyses B—Data derived from a single randomized clinical trial or large non-­ randomized studies C—Consensus of opinion of the experts and/or small studies, retrospective studies, registries

of the high risk of device dislodgement. Ideal RVOT/pulmonary trunk shapes comprise conduits with parallel borders or conduits with a narrowing in the midportion since this provides a safe landing zone for the stent [39].

CMR performed before and within 1  month of the procedure, with analysis of biventricular function and calculation of great vessel blood flow, has shown an improvement in effective RV and LV stroke volume in both patients with predominantly pulmonary stenosis and those with predominantly regurgitation [9, 35]. In patients with predominantly pulmonary stenosis, this is due to decreased RV ESV and improved RV ejection fraction after marked relief of afterload. By contrast, RV ejection fraction remains unchanged in patients with predominantly pulmonary regurgitation, with the improvement in RV and LV effective stroke volume due to abolishment of pulmonary regurgitation [35].

9.7 New Developments 9.7.1 Development of New Percutaneous Devices Because of the wide variation in patient morphology, size, and dynamics of the right ventricular outflow tract (RVOT)/ pulmonary trunk (Fig. 9.9), only ∼15–20% of patients with a hemodynamic and clinical indication for PPVI can be treated with the current device. Thus, 85% of patients with pulmonary dysfunction still require open-heart surgery for treatment. The majority of these patients are those with dilated, dynamic RVOT/pulmonary trunk anatomy (patients with TOF and previous RVOT patches) in whom the current per-

162

a

M. A. Quail et al.

b

c

Fig. 9.12  New percutaneous device for implantation into the dilated outflow tract. (a) Nitinol device, (b) preimplantation in rapid prototyping model, (c) 3D CT reconstruction postimplantation with device in situ in “first-in-man” case

cutaneous device is too small. Over the last decade, several new devices have been developed which can be implanted into the dilated outflow tract either percutaneously with self-­ expandable devices using the Venus p-valve (MedTech, Shanghai, China) [41] or The Harmony™ Transcatheter Pulmonary Valve (Medtronic Inc., Minneapolis, MN, USA) [42, 43] or using a hybrid surgical approach [44]. The Harmony™ device underwent a successful “first-in-man” implantation [42] in 2010. For this procedure, advanced cardiovascular imaging, in combination with patient-specific computer modeling, is crucial to achieve procedural success (Fig. 9.12).

9.7.2 4D Flow CMT in TOF A recent systematic review of 26, 4D flow CMR studies has shown that this technique has huge potential in the assessment of TOF (Fig. 9.13) [45], with use in retrospective flow measurement (with optimization of valve tracking), and assessment of velocity mapping (very time-consuming and often inaccurate on 2D velocity mapping), while presenting novel information for intracardiac kinetic energy quantification, and vortex visualization in both ventricles, outflow tracts, and great vessels. Such information shows promise and may further support the optimization of PVR timing and advanced interventions to treat arrhythmias, though prospec-

tive, randomized, multicentered studies are required to validate these new parameters. The main disadvantages of 4D flow assessment are the long acquisition, reconstruction, and post-processing time, all on which may be supported by new computing and AI algorithms.

9.7.3 The Use of AI in CMR AI offers the potential to change many aspects of congenital heart disease imaging [46]. At present, there are only a few clinically validated examples of AI applications in this field: however, AI has the potential to affect all parts of the patient journey (Fig.  9.14), from improved and automated patient booking, protocolization, and reporting to reduced CMR sequence, reconstruction, and post-processing times, to speed up CMR scanning to support the acquisition of high-­ quality, complex images. Importantly, developments in AI will help move clinical practice to an era of precision patient management, where information from big data (clinical, laboratory, genetic, ECG, imaging, and outcome data) is used to create knowledge that is then drilled down to meet the specific needs of the individual patient in real time. Ultimately, this will improve patient outcomes, patient and clinician experience, and hopefully free up clinician time from many of the mundane processes that we currently need to carry out.

9  Tetralogy of Fallot

a

163

c

b

d

e

f

g

Fig. 9.13  4D flow analysis applications. (a) Manually placed seed positions; (b) particle streamline visualization of the ascending aorta; (c) main pulmonary and branches forward flow; (d) main pulmonary artery regurgitation; (e) flow analysis: net flow across cardiac cycle; (f) 3D vortex core extraction in the right ventricle; (g) 2D visualization of

h

vortices in the right ventricle; (h) kinetic energy mapping from Kanski et  al. JCMR (2015) 17:111 DOI 10.1186/s12968-015-0211-4 used under the terms of the Creative Commons Attribution 4.0 International License (http://creativecommons.org/licenses/by/4.0/)

Fig. 9.14  Aspects of the TOF radiological pathway for CMR imaging where artificial intelligence (AI) or machine learning (ML) may play a role

164

9.8 Discussion In this chapter, we have emphasized the advantages of CMR imaging over other modalities to demonstrate its many strengths in the comprehensive assessment of TOF. However, it is prudent to mention that there are clearly clinical situations in which other modalities are indeed superior and preferable. Echocardiography is a very effective bedside imaging tool and is used as the sole modality for the initial diagnosis of TOF in infancy and indeed the majority of later clinical follow-up. It could rightly be considered that CMR is an adjunct to this primary imaging modality in the management of TOF. In patients who have been treated previously with metallic stents, aneurysm clips, or pacemakers, CT imaging is the preferable imaging modality. CT delineates RVOT and branch pulmonary artery morphology very well, and it has a very rapid acquisition time, which is valuable for critically ill patients. As patients with TOF survive longer, the burden of acquired ischemic heart disease will increase, and it is here that CT imaging of the coronary arteries is superior to CMR. However, the technique cannot quantify vascular flow and, of course, uses ionizing radiation. In our institution, we have observed an increase in the use of cross-sectional imaging modalities with a resultant decrease in diagnostic cardiac catheterizations. Cardiac catheterization is still necessary when therapeutic interventions are required or when intravascular pressures must be discerned. However, the fluoroscopic projections are not suitable for the characterization of complex 3D malformations or ventricular function.

9.8.1 Conclusion The tremendous advance in cross-sectional cardiovascular imaging has changed the landscape in the long-term follow­up and clinical decision-making in TOF. The wealth of reliable, reproducible hemodynamic information provided by CMR studies justifies its recognition as the gold standard. The data provided by CMR in TOF continues to advance our understanding of the disease and will continue to help us manage patients more effectively. Large, multicenter studies such as the INDICATOR Cohort [47] will be necessary to enhance our understanding of the unnatural life course, pathology, and optimal times for intervention in TOF patients. Practical Pearls • TOF is common, and its routine assessment should be familiar to all CMR imagers. • Protocolized follow-up for repaired TOF is important and should include CMR as outlined in the clinical imaging protocol.

M. A. Quail et al.

• Debate remains about the exact timing of PVR in TOF, but an indexed RV end-diastolic volume between 160 and 200 mL/m2, in the presence of severe pulmonary incompetence, seems appropriate. • CMR can be used to select treatment options (watchful waiting, surgical PVR, or percutaneous PVR). • CMR provides excellent longitudinal data to track hemodynamic deterioration, responses to therapy, and changes in myocardial fibrosis.

References 1. Apitz C, Webb GD, Redington AN.  Tetralogy of Fallot. Lancet. 2009;374(9699):1462–71. 2. Murphy JG, Gersh BJ, Mair DD, et  al. Long-term outcome in patients undergoing surgical repair of tetralogy of Fallot. N Engl J Med. 1993;329(9):593–9. 3. Smith CA, McCracken C, Thomas AS, et al. Long-term outcomes of tetralogy of Fallot. A study from the Pediatric cardiac care consortium. JAMA Cardiol. 2019;4(1):34–41. https://doi.org/10.1001/ jamacardio.2018.4255. 4. Chiu SN, Wang JK, Chen HC, et al. Long-term survival and unnatural deaths of patients with repaired tetralogy of Fallot in an Asian cohort. Circ Cardiovasc Qual Outcomes. 2012;5:120–5. https://doi. org/10.1161/CIRCOUTCOMES.111.963603. 5. Park MK. Pediatric cardiology for practitioners. 3rd ed. St. Louis, MI: Mosby; 1996. 6. Suleiman T, Kavinsky CJ, Skerritt C, Kenny D, Ilbawi MN, Caputo M.  Recent development in pulmonary valve replacement after tetralogy of Fallot repair: the emergence of hybrid approaches. Front Surg. 2015;2:22. https://doi.org/10.3389/fsurg.2015.00022. 7. Ghez O, Tsang VT, Frigiola A, et  al. Right ventricular outflow tract reconstruction for pulmonary regurgitation after repair of tetralogy of Fallot. Preliminary results. Eur J Cardiothorac Surg. 2007;31(4):654–8. 8. Bonhoeffer P, Boudjemline Y, Saliba Z, et  al. Percutaneous replacement of pulmonary valve in a right-ventricle to pulmonary-­ artery prosthetic conduit with valve dysfunction. Lancet. 2000;356(9239):1403–5. 9. Khambadkone S, Coats L, Taylor A, et al. Percutaneous pulmonary valve implantation in humans: results in 59 consecutive patients. Circulation. 2005;112(8):1189–97. 10. McElhinney DB, Hellenbrand WE, Zahn EM, et  al. Short- and medium-term outcomes after transcatheter pulmonary valve placement in the expanded multicenter US melody valve trial. Circulation. 2010;122:507–16. https://doi.org/10.1161/ CIRCULATIONAHA.109.921692. 11. Eicken A, Ewert P, Hager A, et  al. Percutaneous pulmonary valve implantation: two-Centre experience with more than 100 patients. Eur Heart J. 2011;32:1260–5. https://doi.org/10.1093/eurheartj/ehq520. 12. Kenny D, Hijazi ZM, Kar S, et al. Percutaneous implantation of the Edwards SAPIEN transcatheter heart valve for conduit failure in the pulmonary position: early phase 1 results from an international multicenter clinical trial. J Am Coll Cardiol. 2011;58:2248–56. https://doi.org/10.1016/j.jacc.2011.07.040. 13. Lurz P, Nordmeyer J, Muthurangu V, et  al. Comparison of bare metal stenting and percutaneous pulmonary valve implantation for treatment of right ventricular outflow tract obstruction: use of an x-ray/magnetic resonance hybrid laboratory for acute physiological assessment. Circulation. 2009;119(23):2995–3001. 14. Van Dijck I, Budts W, Cools B, et  al. Infective endocarditis of a transcatheter pulmonary valve in comparison with surgical

9  Tetralogy of Fallot implants. Heart. 2015;101(10):788–93. https://doi.org/10.1136/ heartjnl-­2014-­306761. 15. Sorensen TS, Korperich H, Greil GF, et al. Operator-independent isotropic three-dimensional magnetic resonance imaging for morphology in congenital heart disease: a validation study. Circulation. 2004;110(2):163–9. 16. Schievano S, Coats L, Migliavacca F, et  al. Variations in right ventricular outflow tract morphology following repair of congenital heart disease: implications for percutaneous pulmonary valve implantation. J Cardiovasc Magn Reson. 2007;9(4):687–95. 17. Geva T, Greil GF, Marshall AC, Landzberg M, Powell AJ.  Gadolinium- enhanced 3-dimensional magnetic resonance angiography of pulmonary blood supply in patients with complex pulmonary stenosis or atresia: comparison with x-ray angiography. Circulation. 2002;106(4):473–8. 18. Barkhausen J, Ruehm SG, Goyen M, Buck T, Laub G, Debatin JF.  MR evaluation of ventricular function: true fast imaging with steady-state precession versus fast low-angle shot cine MR imaging: feasibility study. Radiology. 2001;219(1):264–9. 19. Winter MM, Bernink FJ, Groenink M, et al. Evaluating the systemic right ventricle by CMR: the importance of consistent and reproducible delineation of the cavity. J Cardiovasc Magn Reson. 2008;10:40. 20. Taylor A, Dymarkowski S, Bogaert J.  Valvular heart disease. In: Bogaert J, Dymarkowski S, Taylor A, Muthurangu V, editors. Clinical cardiac MRI.  Heidelberg: Springer; 2012. p.  465–509. ISBN: 978-3-642-23035-6. 21. Nordmeyer J, Gaudin R, Tann OR, et al. MRI may be sufficient for noninvasive assessment of great vessel stents: an in vitro comparison of MRI, CT, and conventional angiography. Am J Roentgenol. 2010;195(4):865–71. 22. Babu-Narayan SV, Kilner PJ, Li W, et al. Ventricular fibrosis suggested by cardiovascular magnetic resonance in adults with repaired tetralogy of Fallot and its relationship to adverse markers of clinical outcome. Circulation. 2006;113(3):405–13. 23. Yiğit H, Ergün E, Öztekin PS, Koşar PN.  Can T1 mapping be an alternative of post-contrast magnetic resonance sequences in patients with surgically corrected tetralogy of Fallot? Anatol J Cardiol. 2020;24(6):377–81. 24. Al-Wakeel-Marquard N, Ferreira da Silva T, Jeuthe S, et  al. Measuring myocardial extracellular volume of the right ventricle in patients with congenital heart disease. Sci Rep. 2021;11(1):2679. https://doi.org/10.1038/s41598-­021-­81440-­z. 25. van den Bosch E, Cuypers JAAE, et  al. Ventricular response to dobutamine stress cardiac magnetic resonance imaging is associated with adverse outcome during 8-year follow-up in patients with repaired tetralogy of Fallot. Eur Heart J Cardiovasc Imaging. 2020;21(9):1039–46. https://doi.org/10.1093/ehjci/jez241. 26. Lurz P, Muthurangu V, Schievano S, et al. Feasibility and reproducibility of biventricular volumetric assessment of cardiac function during exercise using real- time radial k-t SENSE magnetic resonance imaging. J Magn Reson Imaging. 2009;29(5):1062–70. 27. Lurz P, Puranik R, Nordmeyer J, et  al. Improvement in left ventricular filling properties after relief of right ventricle to pulmonary artery conduit obstruction: contribution of septal motion and interventricular mechanical delay. Eur Heart J. 2009;30(18):2266–74. 28. Greutmann M. Tetralogy of Fallot, pulmonary valve replacement, and right ventricular volumes: are we chasing the right target? Eur Heart J. 2016;37:836–9. https://doi.org/10.1093/eurheartj/ehv634. 29. Frigiola A, Tsang V, Bull C, et al. Biventricular response after pulmonary valve replacement for right ventricular outflow tract dysfunction: is age a predictor of outcome? Circulation. 2008;118(14 Suppl):S182–90. 30. Oosterhof T, van Straten A, Vliegen HW, et al. Preoperative thresholds for pulmonary valve replacement in patients with corrected tetralogy of Fallot using cardiovascular magnetic resonance. Circulation. 2007;116(5):545–51. 31. Therrien J, Provost Y, Merchant N, Williams W, Colman J, Webb G. Optimal timing for pulmonary valve replacement in adults after tetralogy of Fallot repair. Am J Cardiol. 2005;95(6):779–82.

165 32. Buechel ER, Dave HH, Kellenberger CJ, et al. Remodelling of the right ventricle after early pulmonary valve replacement in children with repaired tetralogy of Fallot: assessment by cardiovascular magnetic resonance. Eur Heart J. 2005;26(24):2721–7. 33. Stout KK, Daniels CJ, Aboulhosn JA, et  al. 2018 AHA/ACC guideline for the management of adults with congenital heart disease: executive summary: a report of the American College of Cardiology/American Heart Association task force on Clinical practice guidelines. Circulation. 2019;139:e637–97. 34. Baumgartner H, De Backer J, Babu-Narayan SV, et al. ESC scientific document group. 2020 ESC guidelines for the management of adult congenital heart disease. Eur Heart J. 2021;42(6):563–645. https://doi.org/10.1093/eurheartj/ehaa554. 35. Coats L, Khambadkone S, Derrick G, et  al. Physiological consequences of percutaneous pulmonary valve implantation: the different behaviour of volume- and pressure-overloaded ventricles. Eur Heart J. 2007;28(15):1886–93. 36. Mohamed I, Stamm R, Keenan R, Boris Lowe B, Coffey S.  Assessment of disease progression in patients with repaired tetralogy of Fallot using cardiac magnetic resonance imaging: a systematic review. Heart Lung Circ. 2020;29(11):1613–20. https:// doi.org/10.1016/j.hlc.2020.04.017. 37. Secchi F, Lastella G, Monti CB, Barbaro U, Capra D, Zanardo M, Sardanelli F. Late gadolinium enhancement in patients with tetralogy of Fallot: a systematic review. Eur J Radiol. 2021;136:109521. https://doi.org/10.1016/j.ejrad.2021.109521. 38. Ghonim S, Gatzoulis MA, Ernst S, et  al. Predicting survival in repaired tetralogy of Fallot. A lesion-specific and personalized approach. JACC Cardiovasc Imaging. 2022;15(2):257–68. https:// doi.org/10.1016/j.jcmg.2021.07.026. 39. Schievano S, Migliavacca F, Coats L, et al. Percutaneous pulmonary valve implantation based on rapid prototyping of right ventricular outflow tract and pulmonary trunk from MR data. Radiology. 2007;242(2):490–7. 40. Sridharan S, Coats L, Khambadkone S, Taylor AM, Bonhoeffer P. Images in cardiovascular medicine. Transcatheter right ventricular outflow tract intervention: the risk to the coronary circulation. Circulation. 2006;113(25):e934–5. 41. Cao QL, Kenny D, Zhou D, et  al. Early clinical experience with a novel self-expanding percutaneous stent-valve in the native right ventricular outflow tract. Catheter Cardiovasc Interv. 2014;84:1131–7. https://doi.org/10.1002/ccd.25544. 42. Schievano S, Taylor AM, Capelli C, et al. First-in-man implantation of a novel percutaneous valve: a new approach to medical device development. EuroIntervention. 2010;5(6):745–50. 43. Gillespie MJ, Benson LN, Bergersen L, et  al. Patient selection process for the harmony transcatheter pulmonary valve early feasibility study. Am J Cardiol. 2017;120(8):1387–92. https://doi. org/10.1016/j.amjcard.2017.07.034. 44. Chen Q, Turner M, Caputo M, Stoica S, Marianeschi S, Parry A. Pulmonary valve implantation using self-expanding tissue valve without cardiopulmonary bypass reduces operation time and blood product use. J Thorac Cardiovasc Surg. 2013;145:1040–5. https:// doi.org/10.1016/j.jtcvs.2012.05.036. 45. Elsayed A, Gilbert K, Scadeng M, Cowan BR, Pushparajah K, Young AA.  Four-dimensional flow cardiovascular magnetic resonance in tetralogy of Fallot: a systematic review. J Cardiovasc Magn Reson. 2021;23(1):59. https://doi.org/10.1186/ s12968-­021-­00745-­0. 46. Taylor AM.  The role of artificial intelligence in paediatric cardiovascular magnetic resonance imaging. Pediatr Radiol. 2021;52(11):2131–8. https://doi.org/10.1007/s00247-­021-­05218-­1. 47. Geva T, Mulder B, Gauvreau K, et  al. Preoperative predictors of death and sustained ventricular tachycardia after pulmonary valve replacement in patients with repaired tetralogy of Fallot enrolled in the INDICATOR cohort. Circulation. 2018;138(19):2106–15. https://doi.org/10.1161/CIRCULATIONAHA.118.034740.

Ebstein’s Anomaly and Other Tricuspid Valve Anomalies

10

Steve W. Leung and Mushabbar A. Syed

10.1 Introduction

10.2 Tricuspid Valve Anatomy

Tricuspid valve is affected by a wide variety of abnormalities both congenital and acquired. The clinical impact of these abnormalities also varies widely depending upon the lesion and its impact on right heart function. Congenital anomalies of the tricuspid valve are rare and include:

The tricuspid valve is the largest of the four cardiac valves and is located between the right atrium and right ventricle. The tricuspid valve complex consists of leaflets, the fibrous tricuspid annulus, the chordae tendinae, papillary muscles, and the right atrial and right ventricular (RV) myocardium. The tricuspid annulus is an asymmetric, saddle-shaped ellipsoid that is dynamic in nature and can change markedly with loading conditions and during cardiac cycle, e.g., 30% decrease in annular area during systole [1]. Leaflets are usually semicircular or triangular and are attached to the annulus at their base. Typically, tricuspid valve consists of three leaflets: anterior or superior, posterior or inferior, and septal with anterior leaflet being the largest and most mobile while septal being the smallest and least mobile. Septal leaflet lies against the septum and is apically displaced from the mitral annulus. Variations in the structural orientation of tricuspid leaflets and occurrence of accessary leaflets have been reported with one study of 36 adult hearts showed that the number of leaflets can vary from routine three to as many seven [2]. The anterior papillary muscle provides chordae to the anterior and posterior leaflets, and the medial papillary muscle provides chordae to the posterior and septal leaflets. The septal wall gives chordae to the anterior and septal leaflets without the presence of a formal septal papillary muscle.

• • • •

Ebstein’s anomaly Tricuspid atresia Tricuspid stenosis/hypoplasia Tricuspid regurgitation

Among all congenital tricuspid valve defects, tricuspid atresia is more common and is usually diagnosed at or soon after birth. Ebstein’s anomaly is less common, and due to its variable, clinical presentation may not be diagnosed until adulthood.

Supplementary Information The online version contains supplementary material available at https://doi.org/10.1007/978-­3-­031-­29235-­4_10.

10.3 Ebstein’s Anomaly S. W. Leung Division of Cardiovascular Disease, Department of Medicine and Radiology, University of Kentucky, Lexington, KY, USA e-mail: [email protected] M. A. Syed (*) Cardiovascular Imaging, Division of Cardiology, Loyola University Medical Center, Maywood, IL, USA e-mail: [email protected]

Ebstein’s anomaly is a rare congenital heart disease that affects approximately 1 in 200,000 live births and 20 mm or ≥8 mm/m2 indexed to body surface area relative to the mitral valve in the apical four-chamber view is considered diagnostic [6].

10.5 Morphology Ebstein’s anomaly is characterized by the congenital malformations of the right ventricle and tricuspid valve. These malformations include adherence of the septal and posterior leaflets to the myocardium, apical displacement of the tricuspid annulus, redundancy/fenestration and tethering of the anterior leaflet, dilation of the atrialized portion of the right ventricle, and dilation of the right atrioventricular (AV) junction [7]. The posterior leaflet is the most frequently affected leaflet, followed by the septal and anterior leaflets. Depending on the severity of the tethering, patients can present as a neonate or late into adult life. Due to the developmental malformation of the tricuspid valve, patients can

S. W. Leung and M. A. Syed

have tricuspid regurgitation or occasionally tricuspid stenosis. The amount of tethering also determines whether the valve can be repaired or replaced by the Carpentier classification [8]. • Type A: right ventricular volume is adequate. • Type B: large atrialized right ventricle with mobile anterior leaflet of the tricuspid valve. • Type C: severely restricted anterior leaflet of the tricuspid valve. • Type D: almost complete atrialization of the right ventricle except for a small infundibular component. Patients with Type A, B, and C leaflets are likely to benefit from surgical repair. Type D patients require valve replacement.

10.6 Associated Anomalies The most common anomaly associated with Ebstein’s anomaly is interatrial connection such as atrial septal defects and patent foramen ovale, which occurs in upwards of 80% of patients with Ebstein’s anomaly. There are numerous other abnormalities that are associated with Ebstein’s anomaly aside from atrial septal defects (Table 10.1). Ebstein’s anomaly, generally an isolated process, has been described in patients with Down’s syndrome, Marfan syndrome, Noonan syndrome, and left ventricular (LV) non-compaction. Table 10.1  Anatomical anomalies associated with Ebstein’s anomaly Septal defects Atrial septal defect Patent foramen ovale Ventricular septal defect Aortic valve/aorta Bicuspid or atretic aortic valve Aortic coarctation Subaortic stenosis Corrected transposition of the great arteries Patent ductus arteriosus Pulmonic valve/pulmonary artery Pulmonary stenosis Pulmonary atresia Hypoplastic pulmonary arteries Hypertensive pulmonary vascular disease Mitral valve Parachute mitral valve Cleft anterior leaflet of the mitral valve Mitral valve prolapse Other Left ventricular outflow obstruction Tetralogy of Fallot Left ventricular non-compaction

10  Ebstein’s Anomaly and Other Tricuspid Valve Anomalies

10.7 Clinical Presentation Due to the wide spectrum of the severity of the tricuspid malformation, clinical presentation can range from intrauterine fetal demise to incidental finding in an asymptomatic adult patient. Patients who present during their first year of life usually have more severe cardiac disease and present with severe heart failure and cyanosis. In contrast, children and adults more often present with incidental murmurs and arrhythmias such as Wolff–Parkinson–White syndrome [9]. Symptomatic patients may present with right ventricular failure from worsening tricuspid regurgitation and atrial arrhythmia. Common presenting symptoms include: • • • • •

169

CMR can provide great visualization of great vessel abnormalities and quantify intracardiac shunt ratio. Per the most recent ACC/AHA guidelines, serial CMR is recommended to follow these patients. The interval varies based on the patient’s physiologic stage of disease. Physiologic Stages A–D Frequency of testing TTE CMR/CCT

Stage A (mo) 12–24 60

Stage B (mo) 12–24 36

Stage C (mo) 12 24–36

Stage D (mo) 12 12–24

Physiological stage A NYHA I symptoms No hemodynamic or anatomic sequelae No arrhythmias Normal exercise capacity Normal renal/hepatic/pulmonary function B NYHA II symptoms Mild hemodynamic sequelae (mild aortic enlargement, mild ventricular enlargement, mild ventricular dysfunction) Mild valvular disease Trivial or small shunt (not hemodynamically significant) Arrhythmia not requiring treatment Abnormal objective cardiac limitation to exercise C NYHA III symptoms Significant (moderate or greater) valvular disease; moderate or greater ventricular dysfunction (systemic, pulmonic, or both) Moderate aortic enlargement Venous or arterial stenosis Mild or moderate hypoxemia/cyanosis Hemodynamically significant shunt Arrhythmias controlled with treatment Pulmonary hypertension (less than severe) End-organ dysfunction responsive to therapy D NYHA IV symptoms Severe aortic enlargement Arrhythmias refractory to treatment Severe hypoxemia (almost always associated with cyanosis) Severe pulmonary hypertension Eisenmenger syndrome Refractory end-organ dysfunction

Dyspnea on exertion Fatigue and lack of appetite Abdominal distention due to ascites Peripheral edema Palpitations

Due to the high prevalence of septal defects, patients may present later in life with pulmonary hypertension, paradoxical emboli, and Eisenmenger physiology. The electrocardiogram in patients with Ebstein’s anomaly often has interventricular conduction delay or right bundle branch block. The P-waves may indicate marked right atrial enlargement often described as Himalayan P-waves. Since there is a higher prevalence of pre-excitation in this population, delta waves can be seen in some patient’s ECGs. Due to the atrialization of the right ventricle, right ventricular hypertrophy is rare and if present other associated anomalies should be considered. Most cases of Ebstein’s anomaly discovered have been sporadic, but there have been some familial cases. Neonates whose mothers were exposed to lithium have been reported to develop Ebstein’s anomaly.

Adapted from 2018 ACC/AHA Guideline for the Management of Adults with Congenital Heart Disease [10]

a

10.8 Cardiac Magnetic Resonance Imaging 10.8.1 Indications

10.8.2 Goals of Imaging

Echocardiography has been the main tool in the diagnosis of Ebstein’s anomaly. However, in patients with poor echocardiographic windows, CMR can be an alternative method in making the diagnosis. In patients with known Ebstein’s anomaly, CMR can provide valuable information in quantifying right ventricular size and function, along with severity of tricuspid regurgitation. Serial CMR can be performed to serially monitor right ventricular size and function, as well as severity of tricuspid regurgitation. Since patients with Ebstein’s anomaly often have other associated anomalies,

1. To determine the extent of apical displacement of the tricuspid valve and assess mobility of the anterior leaflet (Carpentier classification) 2. To accurately assess right ventricular size and function 3. To identify tricuspid regurgitation and quantify severity 4. To identify associated lesions such as atrial septal defect, ventricular septal defect, and calculate the amount of shunt (Qp:Qs ratio) 5. To determine aortic or pulmonary artery anomalies 6. To identify right ventricular outflow tract obstructions

S. W. Leung and M. A. Syed

170

10.8.3 CMR Sequences and Imaging Protocols A sample CMR scanning protocol is listed in Table 10.2.

10.8.3.1 Scout Imaging Scout imaging in various planes can help quickly identify any significant right ventricular enlargement and any apparent associated lesions in the aorta or pulmonary artery such as coarctation or hypoplastic pulmonary artery. Since most patients who present have enlarged right heart, the cardiac anatomy is likely to be distorted. The scout images can then help with accurate prescription of the long- and short-axis cines and ensure adequate coverage of the ventricles for quantification of size and function. 10.8.3.2 Cine Imaging—Diagnosis, Chamber Size and Function, and Septal Defects For the diagnosis of Ebstein’s anomaly, four-chamber steady-­ state free precession (SSFP) cine can identify the extent of apical displacement of the tricuspid valve compared to the mitral valve (Fig. 10.1 and Movie 10.2). The diagnosis can be made if the apical displacement of the tricuspid valve compared to the mitral valve is >20  mm or ≥8  mm/m2 indexed to body surface area. The identification of the tricuspid leaflets and the amount of tethering can be useful in determining type of surgical treatment. However, due to the complex shape and thin tricuspid valve leaflet and relatively limited spatial resolution of CMR, it may be difficult to visualize leaflet adherence to the myocardium. The short-axis cines from base to apex are routinely used for calculating LV and RV size and function (Fig. 10.2) [11]. Alternatively, axial plane cines can also be used for calculating RV size and function; however, the method of acquiring RV size and function in serial follow-up studies should be consistent with the first study [12]. To identify septal defects, four-chamber stack of cines through the atrial and ventricular septum extending from the superior vena cava to the level of atrioventricular (AV) valves should be obtained for interatrial or interventricular connections (Fig. 10.3). 10.8.3.3 In Plane Phase Contrast Imaging: Evaluation of Abnormal Flow SSFP cine imaging can often miss valvular regurgitation or flow through septal defects as the sequence is designed to suppress flow for better delineation of endocardial borders. Therefore, ECG-gated, free-breathing, in-plane phase contrast imaging can be helpful in clearly identifying any tricuspid regurgitation or flow through septal defects in patients with Ebstein’s anomaly. Qualitatively, tricuspid regurgitation can be evaluated by in-plane phase contrast imaging in the four-chamber view or right atrial/ventricular two-chamber view with the axis parallel to the predicted tricuspid regurgi-

Table 10.2  Imaging protocol example •  Scout (localizers) scan: in three-plane (axial, sagittal, and coronal), axial stack, four-chamber, short-axis plane •  SSFP cine images: Short-axis, multislice stack from base to apex Three-chamber, two-chamber, four-chamber views Two-chamber of the right atrium/right ventricle LVOT and RVOT Four-chamber view, stack for full volume coverage (for identification of septal defects) Optional: axial plane, cine stack from top to bottom of the right ventricle •  First pass contrast-enhanced MRA of pulmonary arteries and aorta •  In-plane phase contrast flow imaging (ECG-gated, free-breathing) Four-chamber view LVOT RVOT Optional: two-chamber view of the right atrium/ventricle • Through-plane phase contrast imaging (ECG-gated, free-­ breathing) of the main pulmonary artery (Qp) and ascending aorta (Qs) • Late gadolinium enhancement imaging in short-axis stack, two-chamber, three-chamber, four-chamber views (usually 10–20 min post-contrast) •O  ptional: through-plane phase contrast of static gel phantom with identical acquisition parameters as Qp and Qs

Fig. 10.1  Still frame in an SSFP four-chamber view shows the apical displacement (double arrow) of the tricuspid valve compared to the mitral valve. End-diastolic frame is used for measuring the apical displacement

tant jet (Movie 10.3). The velocity encoding (VENC) setting should be at least 150 cm/s. For septal defects, the appropriate four-chamber view of the suspected area of the defect should be evaluated with in-plane phase contrast. Due to slow flow in the atrium, the VENC setting is usually set at 50–100 cm/s. Since patients with Ebstein’s anomaly can also have left or right ventricular outflow tract (LVOT or RVOT)

10  Ebstein’s Anomaly and Other Tricuspid Valve Anomalies

Fig. 10.2  Still frame of short-axis stack for evaluation of ventricular size and function

171

172

S. W. Leung and M. A. Syed

Fig. 10.3  Still frame of four-chamber stack for evaluation of septal defects

obstruction, LVOT and RVOT in-plane phase contrast can help identify any significant turbulent flow suggestive of obstruction across the outflow tracts (Fig.  10.4 and Movie 10.4). Occasionally, patients with Ebstein’s anomaly have tricuspid stenosis. Severe tricuspid stenosis has been identified as a right atrium to right ventricle gradient of >5 mmHg during diastole by Doppler echocardiography. In-plane phase contrast imaging in a four-chamber or right-sided two-­ chamber orientation can identify increased velocity through the tricuspid valve during diastole as a sign of tricuspid stenosis. Initial VENC of 100  cm/s should be used. If the jet aliases during diastole, then the gradient across the tricuspid valve is >4 mmHg by using the simplified Bernoulli’s equation (gradient = 4v2). Be aware that in patients with significant tricuspid regurgitation that there can be increased

velocities or gradients due to increase flow, and not due to anatomic tricuspid stenosis.

10.8.3.4 Through-Plane Phase Contrast: Quantification of Tricuspid Regurgitation Patients with Ebstein’s anomaly often develop significant tricuspid regurgitation. In echocardiography, severe tricuspid regurgitation is determined by color jet size, vena contracta width >0.7 cm2, and systolic reversal of flow in the hepatic veins. In CMR, there have not been established criteria for severe tricuspid regurgitation; however, in general, valvular regurgitant fraction >40% is considered severe. Unlike the aortic and pulmonic valves, the tricuspid annulus goes through significantly complex motion during systole, which renders direct measurements by through-plane phase con-

10  Ebstein’s Anomaly and Other Tricuspid Valve Anomalies

173

volume (RVESV) from the right ventricular end-diastolic volume (RVEDV). RVESV and RVEDV are obtained from either the short-axis stack or axial stack SSFP cine images. Tricuspid regurgitant volume (RV TR) can then be calculated by:

RVTR =Qp − ( RVEDV − RVESV ) =Qp − RVSV. Regurgitant fraction (RF) can be obtained by:



RF% = ( RVTR / RVSV ) ×100%.

This method can only be used if there is no interventricular shunt. Without phase contrast flow measurements, RVTR can also be calculated by: RV =RVSV − ( LVEDV − LVESV ) =RVSV − LVSV. TR This can only be performed if there are no other significant valvular regurgitations (aortic, mitral, or pulmonic) or interventricular shunt.

10.8.3.5 Through-Plane Phase Contrast: Quantification of Shunt Ratio Since patients with Ebstein’s anomaly frequently have associated atrial or ventricular septal defects, shunt ratio quantification (Qp/Qs) can be helpful in determining need for closure of these defects. Shunt ratio quantification should be obtained from through-plane phase contrast. Qp prescription has been described above. Through-plane Fig. 10.4  Still frame of right ventricular outflow tract SSFP cine for phase contrast of an appropriate axial slice above the aortic evaluation of right ventricular outflow tract obstruction valve can be obtained for effective systemic flow (Qs) (Fig. 10.6). The VENC for flow measurement in ascending trast CMR of the tricuspid valve difficult. For best results, a aorta is usually set at 200 cm/s. short-axis plane can be prescribed using four-chamber cine With phase contrast imaging, flow acquisition can be SSFP to align the imaging plane with the valve annulus at erroneous due to background phase offset errors from non-­ end-systole. Additionally, the tricuspid regurgitant volume compensated eddy currents [13]. This error can be corrected and regurgitant fraction can be measured indirectly from by obtaining phase contrast images from static gel phantoms effective pulmonic flow and right ventricular stroke volume. post-acquisition with heart rate simulator simulating a heart Effective pulmonary flow (Qp) can be obtained by ECG-­ rate similar to the time of Qp and Qs acquisition or backgated, free-breathing, through-plane phase contrast imaging ground phase correction based on a regional of interest in the of the main pulmonary artery cross section above the pul- static chest wall. The correction should then be used to adjust monic valve. Often, this prescription can be obtained from a the original data. It’s important to note that the table position perpendicular slice from the right ventricular outflow tract should not be reset prior to obtaining the phantom images, as view (sagittal plane) above the pulmonic valve (Fig.  10.5) this can cause positioning errors in finding the original flow starting at VENC of 150 cm/s. The resultant velocity images positions and acquisition of the correct background phase should be immediately assessed for aliasing (low VENC set- offset errors. ting) and repeated if necessary by increasing the VENC setting. Also check the magnitude images for appropriate vessel shape (arteries should be round) and phase wrap artifacts. 10.8.4 Contrast-Enhanced Magnetic Phase wrap does not significantly affect the precision of the Resonance Angiography measurements as long as wrap is not superimposed on the vessel of interest. Right ventricular stroke volume (RVSV) is Patients with Ebstein’s anomaly can have associated pulmoobtained by subtracting the right ventricular end-systolic nary artery or aortic anomalies. Contrast-enhanced magnetic

174

a

S. W. Leung and M. A. Syed

b

Fig. 10.5  Still frame of the main pulmonary artery just above the pulmonic valve for effective pulmonary flow (Qp) quantification. (a) Magnitude image for anatomy and (b) is corresponding phase image. Vessel of interest should be circular to quantify flow

a

b

Fig. 10.6  Still frame of the ascending aorta at pulmonary artery bifurcation for effective systemic flow (Qs) quantification. (a) Magnitude image for anatomy and (b) is corresponding phase image. Vessel of interest should be circular to quantify flow

10  Ebstein’s Anomaly and Other Tricuspid Valve Anomalies

resonance angiography (MRA) can provide a three-­ dimensional view to detect hypoplastic pulmonary arteries, coarctation of the aorta, and transposition of the great arteries. It can also provide accurate vascular measurements for planning surgical treatment of these anomalies.

10.8.5 Late Gadolinium Enhancement Imaging Late gadolinium enhancement imaging has been quite useful in detecting scar or fibrosis in patients with myocardial infarction and various cardiomyopathies. In patients with Ebstein’s anomaly, late gadolinium enhancement can be used to identify fibrotic changes of the atrialized ventricular wall [14]. RV septal late gadolinium enhancement and elevated LV extracellular volume in patients with Ebstein’s anomaly are associated with worse LV and RV function and NYHA class [15]. In patients with tricuspid stenosis or right ventricular dysfunction, late gadolinium imaging can also identify right atrial and right ventricular thrombus.

10.9 CMR in Comparison with Other Imaging Modalities Echocardiography remains the primary method of diagnosing Ebstein’s anomaly and identifying tricuspid leaflet deformities for possible surgical correction. Due to the higher spatial resolution, the tethering of the tricuspid leaflet can be more easily identified. Doppler echocardiography can also be helpful in evaluation of tricuspid regurgitation severity. However, in cases where there are limited acoustic windows and/or poor spatial resolution in deep structures, CMR is helpful in providing essential information. Since structures are more clearly defined and volumetric measurements are more reliable and reproducible than echocardiography, CMR is important in serial imaging to follow patients with established diagnosis of Ebstein’s anomaly. Since patients with Ebstein’s anomaly may have various associated vascular anomalies, CMR can help identify these other findings that may have been missed on echocardiography. With the development of multi-slice computed tomography (CT) with electrocardiographic gating, CT scan has also become useful in evaluation of congenital heart disease. Since cardiac CT can provide excellent spatial resolution, complete volumetric coverage, and functional analysis, it can be a valuable tool when echocardiography and CMR cannot be performed adequately. This can be due to inadequate acoustic windows or suboptimal right ventricular visualization for echocardiography, contraindication to CMR due to metallic implants, patients unable to lay flat for pro-

175

longed periods of time, or unable to hold their breaths repeatedly. However, at this time, cardiac CT cannot adequately quantify the amount of tricuspid regurgitation unless there is no other concomitant valvular regurgitation or intracardiac shunts. Since cardiac function evaluation with CT requires high doses of radiation, this should be limited to as the last resort especially in young patients or women of child-­bearing age.

10.10 Limitations and Common Pitfalls For patients with Ebstein’s anomaly who are referred for CMR imaging, prior imaging studies should be reviewed to develop a plan of scanning to reduce scanning time. One of the limitations of CMR is that image acquisition can take a prolonged period of time. Some patients are unable to tolerate such prolonged scans due to claustrophobia, fatigue, back pain, or other issues. If a patient has known atrial septal defect that has been adequately evaluated by echocardiography, one can consider skipping the four-chamber cine stack. If a patient does not have RVOT or LVOT obstruction on Doppler echocardiography, one can consider skipping these imaging planes to reduce acquisition time. Other limitations include patient cooperation in following breathing commands, which can limit the clarity in the SSFP cine images and late gadolinium enhancement images. The spatial resolution limits visualization of the tricuspid valve leaflet as clearly as echocardiography. Despite the common use of phase contrast for flow measurements, it is important to perform background offset error correction to obtain more accurate flow quantification data.

10.11 Tricuspid Atresia Tricuspid atresia is a form of cyanotic congenital heart disease characterized by the complete agenesis of the tricuspid valve. In a 2019 study, it was estimated that about 404 babies in the United States are born with tricuspid atresia with an incidence of 1 in every 9751 live births [16]. In the current era, the majority of patients are diagnosed during pregnancy or soon after birth with echocardiogram as the first-line imaging modality. In most patients, the tricuspid inlet appears as a dimple in the right atrium, and in rare form, there is fusion of partial delaminated leaflets and the formation of membrane (Ebstein type). Tricuspid atresia is associated with hypoplastic right ventricle, atrial septal defect, ventricular septal defect, and pulmonary obstruction. This lesion has been classified based on the relationship of the great arteries, presence of a VSD and degree of pulmonary obstruction [17].

176

S. W. Leung and M. A. Syed

Type I (70–80%): Normal anatomy of the great arteries Subgroup a—Intact ventricular septum with pulmonary atresia Subgroup b—Small VSD with pulmonary stenosis (PS)/ hypoplasia Subgroup c—Large VSD without PS Type II (12–25%): D-transposition of the great arteries (D-TGA) Subgroup a—VSD with pulmonary atresia Subgroup b—VSD with PS or hypoplasia Subgroup c—VSD without PS Type III (3–6%): Malposition defects of the great arteries other than D-TGA (e.g., truncus arteriosus, atrioventricular septal defects, and double outlet right ventricle) Echocardiography is diagnostic for tricuspid atresia with limited role for other imaging modalities. CMR is primarily used in the follow-up of these patients after Fontan surgery and is discussed in detail in the Single Ventricle and Fontan Procedures chapter.

10.12 Other Tricuspid Valve Anomalies Isolated congenital tricuspid stenosis is an extremely rare congenital heart malformation. The condition is a result of hypoplasia and thickening of the tricuspid valve, deformity of the chordae, or malformation of the entire subvalvular apparatus including parachute valve [18]. It is commonly associated with patent foramen ovale or atrial septal defect [19–21]. They are also associated with right-sided hypoplasia including right ventricular hypoplasia and pulmonic stenosis or atresia. SSFP cine in the long-axis views of the right atrium/right ventricle can be helpful in identifying the cause of tricuspid stenosis (Fig. 10.7 and Movie 10.5). The tricuspid gradient can be identified by ECG-gated, free-breathing, in-plane phase contrast imaging in the long axis through the tricuspid valve as described above. Due to the tricuspid annular plane systolic excursion, multiple cross sections through the tricuspid valve with no skip may be required to obtain a planimetry of the valve opening and optimal through-plane phase contrast to obtain the maximal gradient through the valve. A short-axis plane using four-chamber cine SSFP to align the imaging plane with the valve annulus at end-systole usually provides the best result for gradient estimation. Congenital tricuspid regurgitation is generally due to Ebstein’s anomaly. Isolated congenital tricuspid regurgitation is even less common. The valvular insufficiency can be caused by tricuspid leaflet prolapse, hypoplastic or cleft leaflets, absence of papillary muscle or chordae, or annular dilation [21–24]. Functional causes include right ventricular outflow tract obstruction or right ventricular dysfunction [18].

Fig. 10.7  Still frame of four-chamber SSFP cine demonstrating small tricuspid valve annulus. There is small pericardial effusion (asterisk) and prominent intrapericardial fat (arrows)

In these patients, CMR can be used to calculate regurgitant volume and fraction similar to the methods mentioned in Ebstein’s anomaly. SSFP cine imaging of the tricuspid valve in long-axis views can be helpful in identifying tricuspid valve prolapse and other causes of tricuspid regurgitation. Right ventricular volumes can also be followed over time. The role of CMR in these cases is similar to patients with Ebstein’s anomaly. In patients who do not have optimal acoustic windows for echocardiography, the need to evaluate other associated anomalies, or serial follow-up of right-sided volumes, CMR can be beneficial.

10.13 Conclusions Ebstein’s anomaly and congenital tricuspid anomalies are rare congenital heart defects that affect less than 1% of all congenital heart diseases. Although echocardiography is the primary modality for diagnosis of these rare anomalies, it is limited in evaluation of the right ventricular size and ­function. Adult patients who present with little to no symptoms would need to be followed for years prior to the necessary surgical correction due to right ventricular enlargement and failure. CMR not only provides the capability to diagnose these rare anomalies but also can provide accurate serial measurements of right ventricular volume and function and severity of tricuspid regurgitation. Many of these patients also present with various associated anomalies that can be detected readily by a comprehensive CMR study. In patients with associated shunts, quantification of pulmonary artery to aorta flow ratio (Qp/Qs) can be performed to assess the severity of these shunts.

10  Ebstein’s Anomaly and Other Tricuspid Valve Anomalies

177

7. Edwards WD.  Embryology and pathologic features of Ebstein’s anomaly. Prog Pediatr Cardiol. 1993;2(1):5–15. 8. Carpentier A, et  al. A new reconstructive operation for Ebstein’s • Review of prior imaging studies (especially echocardioanomaly of the tricuspid valve. J Thorac Cardiovasc Surg. graphic images) is extremely helpful in planning scanning 1988;96(1):92–101. 9. Celermajer DS, et al. Ebstein’s anomaly: presentation and outcome protocol to minimize scan time. from fetus to adult. J Am Coll Cardiol. 1994;23(1):170–6. • Echocardiography is usually the first-line imaging modal10. Stout KK, Daniels CJ, Aboulhosn JA, et al. 2018 AHA/ACC ity for the diagnosis of tricuspid valve anomalies. CMR Guideline for the Management of Adults With Congenital Heart has a complimentary role for the evaluation of right venDisease: A Report of the American College of Cardiology/ American Heart Association Task Force on Clinical Practice tricle and other associated anomalies. Guidelines. Circulation. 2019;139(14):e698–e800 • CMR can be used to evaluate tricuspid valvular regurgita11. Grothues F, et al. Interstudy reproducibility of right ventricular voltion/stenosis, accurate measurements of right and left umes, function, and mass with cardiovascular magnetic resonance. ventricular size and function, associated anomalies such Am Heart J. 2004;147(2):218–23. 12. Alfakih K, et  al. Comparison of right ventricular volume meaas septal defects and great vessel abnormalities, and quansurements between axial and short axis orientation using steady-­ tification of shunt ratio. state free precession magnetic resonance imaging. J Magn Reson • Serial CMR imaging for assessment of right ventricular Imaging. 2003;18(1):25–32. size and function should be performed in the same orien- 13. Holland BJ, Printz BF, Lai WW.  Baseline correction of phase-­ contrast images in congenital cardiovascular magnetic resonance. tation as the initial study (short-axis stack or axial plane J Cardiovasc Magn Reson. 2010;12:11. stack cines). 14. Nakamura I, et al. Ebstein anomaly by cardiac magnetic resonance • Interatrial and interventricular communication can be difimaging. J Am Coll Cardiol. 2009;53(17):1568. ficult to see on SSFP cine images. ECG-gated, free-­ 15. Yang D, Li X, Sun JY, Cheng W, Greiser A, Zhang TJ, Liu H, Wan K, Luo Y, An Q, Chung YC, Han Y, Chen YC. Cardiovascular magbreathing, in-plane phase contrast flow imaging is helpful netic resonance evidence of myocardial fibrosis and its clinical sigin these situations. nificance in adolescent and adult patients with Ebstein’s anomaly. J • 3D contrasted MRA can be helpful in defining any aortic Cardiovasc Magn Reson. 2018;20(1):69. or pulmonary artery abnormalities. 16. Mai CT, Isenburg JL, Canfield MA, et  al. For the National Birth Defects Prevention Network. National population-based estimates • Flow quantification for Qp, Qs, and tricuspid regurgitafor major birth defects, 2010-2014. Birth Defects Res. 2019;111:1– tion measurements should be carefully planned. 16. https://doi.org/10.1002/bdr2.1589. 17. Rao PS.  A unified classification of tricuspid atresia. Am Heart J. 1980;99:799–804. 18. Dearani JA, Danielson GK. Congenital heart surgery nomenclature and database project: Ebstein’s anomaly and tricuspid valve disReferences ease. Ann Thorac Surg. 2000;69(4 Suppl):S106–17. 19. Chuah SY, Hughes-Nurse J, Rowlands DB. A successful pregnancy 1. Yucel E, Bertrand PB, Churchill JL, Namasivayam M. The tricusin a patient with congenital tricuspid stenosis and a patent oval forapid valve in review: anatomy, pathophysiology and echocardiomen. Int J Cardiol. 1992;34(1):112–4. graphic assessment with focus on functional tricuspid regurgitation. 20. Khan AN, Boatman J, Anderson AS.  Management of new-onset J Thorac Dis. 2020;12:2945–54. congestive heart failure in a patient with complex congenital heart 2. Lama P, Tamang BK, Kulkarni J. Morphometry and aberrant mordisease. Congest Heart Fail. 2002;8(1):54–6. phology of the adult human tricuspid valve leaflets. Anat Sci Int. 21. Krishnamoorthy KM. Balloon dilatation of isolated congenital tri2016;91:143–50. cuspid stenosis. Int J Cardiol. 2003;89(1):119–21. 3. Attenhofer Jost CH, et  al. Ebstein’s anomaly. Circulation. 22. Kobza R, et  al. Aberrant tendinous chords with tethering of the 2007;115(2):277–85. tricuspid leaflets: a congenital anomaly causing severe tricuspid 4. Ebstein W.  Ueber einen sehr seltenen fall von insufficienz der regurgitation. Heart. 2004;90(3):319–23. valvula tricuspidatis, bedingt durch eine angeborene hochgradige 23. Motoyoshi N, et al. Cleft on tricuspid anterior leaflet. Ann Thorac Missbildung derselben. Arch Anat Physiol. 1866;33:238–54. Surg. 2001;71(4):1350–1. 5. Soloff LA, Stauffer HM, Zatuchni J. Ebstein’s disease: report of the 24. Katogi T, et al. Surgical management of isolated congenital tricusfirst case diagnosed during life. Am J Med Sci. 1951;222(5):554–61. pid regurgitation. Ann Thorac Surg. 1998;66(5):1571–4. 6. Shiina A, et  al. Two-dimensional echocardiographic spectrum of Ebstein’s anomaly: detailed anatomic assessment. J Am Coll Cardiol. 1984;3(2 Pt 1):356–70.

Practical Pearls

Abnormalities of Left Ventricular Inflow and Outflow

11

Tal Geva and Puja Banka

11.1 Introduction Congenital abnormalities of left ventricular inflow and outflow include abnormalities of the left atrium, mitral valve (supravalvar, valvar, and subvalvar), and abnormalities of the left ventricular outflow tract, the aortic valve, and supravalvar area. Cardiac magnetic resonance imaging (CMR) has become an important adjunctive tool in evaluating and following patients with this group of anomalies. This chapter reviews the role of CMR in the care of patients with congenital abnormalities of left ventricular inflow and outflow. In addition to describing the morphologic abnormalities and their clinical presentations, the indications and limitations of CMR in each condition are discussed, and a suggested CMR examination protocol is provided.

11.2 Abnormalities of Left Ventricular Inflow 11.2.1 Left Atrium 11.2.1.1 Definitions Abnormalities of the left atrium include congenital left atrial aneurysm and cor triatriatum. The former is characterized by intrinsic left atrial enlargement out of proportion to the hemodynamic load on the left atrium. The latter is characterized by a dividing “membrane” within the left atrium resulting in a proximal pulmonary venous chamber and a distal supramitral chamber containing the appendage. 11.2.1.2 Congenital Left Atrial Aneurysm Morphologic and Functional Abnormalities Congenital left atrial aneurysm is a rare anomaly associated with dysplasia of the left atrial myocardium [1]. This anomaly, also called giant left atrium, is characterized by a markedly dilated left atrium with thinning or partial absence of the myocardium. The left atrial enlargement is out of proportion to its hemodynamic load. The diagnosis is made in the absence of an inflammatory or degenerative process [2].

Supplementary Information The online version contains supplementary material available at https://doi.org/10.1007/978-­3-­031-­29235-­4_11. T. Geva (*) Department of Pediatrics, Harvard Medical School, Boston, MA, USA Department of Cardiology, Boston Children’s Hospital, Boston, MA, USA e-mail: [email protected] P. Banka Department of Pediatrics, Harvard Medical School, Boston, MA, USA Department of Cardiology, Boston Children’s Hospital, Boston, MA, USA e-mail: [email protected]

Associated Anomalies The etiology and morphogenesis of congenital left atrial aneurysm are poorly understood. This lesion is generally found as an isolated condition, and the presence of associated cardiac anomalies that can cause left atrial dilation usually excludes it from the differential diagnosis. Clinical Presentation Congenital left atrial aneurysm often presents as an incidental finding on radiographic imaging done for other reasons [3]. In some cases, however, patients present with tachyarrhythmias, cardiac arrest, pericardial tamponade from rupture of the aneurysm, systemic embolization of atrial

© The Author(s), under exclusive license to Springer Nature Switzerland AG 2023 M. A. Syed, R. H. Mohiaddin (eds.), Magnetic Resonance Imaging of Congenital Heart Disease, https://doi.org/10.1007/978-3-031-29235-4_11

179

180

T. Geva and P. Banka

thrombi, respiratory distress, or heart failure [2, 4]. Since the aneurysm can increase in size over time, long-term followup is generally recommended to monitor its size and ­ ­associated complications. Surgical resection may be considered, especially in symptomatic patients [2].

ily reproducible between scans (e.g., axial or coronal planes), or care should be taken to duplicate the plane used on prior examinations. Alternatively, left atrial volume can be calculated using the summation of discs technique, and this can be followed over time.

CMR of Congenital Left Atrial Aneurysm The aneurysm can be visualized by several techniques including transthoracic and transesophageal echocardiogram, computed tomography, and conventional angiography. CMR can be adjunctive to these diagnostic modalities by providing comprehensive visualization of the aneurysm, imaging of adjacent structures, and measurements of the aneurysm volume [2]. Other causes for left atrial enlargement, such as mitral valve abnormalities or intracardiac shunts, can be excluded. In addition, CMR can be used to assess for thrombus within the aneurysm. Indications for CMR in patients with suspected or known congenital left atrial aneurysm include confirmation of the diagnosis, longitudinal assessment of aneurysm size, and evaluation for thrombus within the aneurysm [4]. CMR examination in a patient suspected of having congenital left atrial aneurysm may include the following:

11.2.1.3 Cor Triatriatum

• Electrocardiogram (ECG)-triggered, breath-hold cine steady-state free precession (SSFP) in the following planes: –– Axial and coronal planes through the entirety of the aneurysm to allow measurements of aneurysm size and visualization of potential thrombi –– Left ventricle (LV) two-chamber (vertical long-axis) –– Four-chamber stack (horizontal long-axis) –– LV three-chamber view parallel to the left ventricular outflow –– Ventricular short-axis stack with 12–14  equidistant slices covering the entire length of both ventricles to evaluate ventricular volumes and function and to exclude other causes of left atrial enlargement • ECG-triggered, breath-hold turbo (fast) spin echo sequence with blood suppression for assessment of left atrial wall thickness • ECG-triggered, free-breathing cine phase contrast flow measurements in the proximal ascending aorta, main pulmonary artery, and, in selected cases, across the atrioventricular (AV) valves to assess for shunts • ECG-triggered, breath-hold, phase sensitive late gadolinium enhancement (LGE) imaging with a long inversion time (e.g., 600 ms) for evaluation of possible thrombi When CMR is used to follow patients longitudinally, care must be taken to make measurements of the aneurysm that are comparable from one scan to the next. For linear measurements, therefore, prescribed planes should either be eas-

Morphologic and Functional Abnormalities Cor triatriatum is a rare lesion comprising approximately 0.1–0.4% of congenital heart disease cases. The lesion is considered a form of stenosis of the common pulmonary vein due to a fibromuscular “membrane” or “diaphragm” that divides the left atrium into two chambers: a proximal chamber that receives the pulmonary veins and a distal chamber that communicates with the left atrial appendage and the mitral valve (Fig. 11.1; Movies 11.1 and 11.2) [5]. Several developmental theories for this lesion have been proposed, but the most commonly accepted hypothesis is incomplete incorporation of the common pulmonary vein to the posterior aspect of the developing left atrium [5]. Some variants of cor triatriatum, however, are not consistent with this explanation. Examples include cases in which the proximal chamber receives only some of the pulmonary veins, a condition termed subtotal cor triatriatum. In typical cor triatriatum, the partitioning left atrial membrane forms a windsock, which is directed toward the mitral valve. The opening in the windsock-like membrane varies in size from few millimeters to about 1 cm. The distal chamber that communicates with the left atrial appendage and the mitral valve also has the fossa ovalis lying between it and the right atrium. In most cases, neither the proximal nor the distal chambers communicate with the right atrium. In some cases, the distal (supramitral) chamber communicates with the right atrium through a patent foramen ovale or a secundum atrial septal defect. Rarely, the proximal (pulmonary venous) chamber communicates with the right atrium. In the latter case, the possibility of a sinus venosus defect should be considered. The hemodynamic burden due to cor triatriatum is obstruction to pulmonary venous return causing pulmonary venous and arterial hypertension. From a hemodynamic standpoint, cor triatriatum can be viewed as a form of mitral stenosis. Associated Anomalies Cor triatriatum can be seen in isolation or with a number of other congenital heart defects, including partially anomalous pulmonary venous connection and secundum atrial septal defect [5]. Clinical Presentation The clinical presentation of this lesion is similar to that of mitral stenosis and depends on the size of the opening in the membrane separating the two left atrial chambers. Rarely,

11  Abnormalities of Left Ventricular Inflow and Outflow

a

181

b

Fig. 11.1  Cor triatriatum. (a) Cine SSFP image in a four-chamber plane showing the cor triatriatum membrane dividing the left atrium into two chambers, a proximal chamber that receives the pulmonary veins (LA1) and a distal chamber that communicates with the mitral

valve and left atrial appendage (LA2). (b) In-plane cine phase contrast flow mapping in the four-chamber plane demonstrating flow across the defect in the cor triatriatum membrane (arrows). LV left ventricle, MV mitral valve, RA right atrium, RV right ventricle

the pulmonary venous chamber can decompress through a defect between the proximal chamber and the right atrium. When the communication between the two left atrial chambers is small, patients develop signs and symptoms of pulmonary venous obstruction such as pulmonary edema and hypertension or decreased cardiac output. If, however, the opening is large, symptoms may be minimal or absent. Older patients can also exhibit embolic phenomenon from thrombus formation in the atrium.

• Extended four-chamber covering the entire left atrium (Fig. 11.1a; Movie 11.1) • LV three-chamber view parallel to the left ventricular outflow • Ventricular short-axis stack to evaluate ventricular volumes and function • Gadolinium-enhanced magnetic resonance angiogram (MRA) to assess for anomalous pulmonary venous connections and other associated anomalies • ECG-triggered, free-breathing cine phase contrast flow measurements in the proximal ascending aorta, main pulmonary artery, and, in selected cases, across the AV valves to assess for shunts. In-plane flow velocity mapping in the four-chamber plane can provide additional information regarding the location and size of the opening within the cor triatriatum membrane (Fig.  11.1b; Movie 11.2). • Optional sequences: –– ECG-triggered, breath-hold turbo (fast) spin echo sequence with blood suppression for further imaging of the cor triatriatum membrane –– ECG-triggered, respiratory-navigated, free breathing three-dimensional isotropic SSFP for evaluation of the coronary arteries –– In patients with possible myocardial scar, LGE imaging performed 10–20 min after contrast administration in the following planes: ventricular short-axis, ­four-­chamber, LV two- and three-chamber, and RV two-chamber

CMR of Cor Triatriatum Given its ubiquitous availability, excellent spatial and temporal resolutions, and its ability to evaluate the hemodynamic burden of cor triatriatum by Doppler, echocardiography is the primary imaging modality used in this condition. CMR can be helpful in two circumstances: (1) as a substitute for transesophageal echocardiography in patients with poor acoustic windows when transthoracic echocardiography does not provide adequate data [6] and (2) to evaluate associated defects such as anomalous pulmonary venous connections [7]. CMR can also be used to measure intracardiac shunts, as well as right ventricular size and function. The following CMR examination protocol can be used in patients with cor triatriatum: • ECG-triggered, breath-hold cine SSFP in the following planes: • LV two-chamber (vertical long-axis) • Right ventricle (RV) two-chamber (vertical long-axis)

182

T. Geva and P. Banka

11.2.2 Mitral Valve 11.2.2.1 Definitions The normal mitral valve has a saddle-shaped annulus, two leaflets with a larger anterior and a smaller posterior leaflet, and primary and secondary chordal attachments to two left ventricular papillary muscles. Each of these components of the valve apparatus plays an important role in valve function, and abnormalities at any level of the valve can result in obstruction to inflow (mitral stenosis) or to incompetence (mitral regurgitation). 11.2.2.2 Mitral Stenosis Morphologic and Functional Abnormalities Congenital mitral stenosis comprises a diverse group of valve morphologies, function, and natural history and is often associated with additional left heart obstructive lesions [8, 9]. Common to all forms of congenital mitral stenosis is anatomic abnormality of one or more components of the valve anatomy leading to narrowing of left ventricular inflow. Abnormalities at any level of the valve apparatus can result in obstruction to left ventricular inflow. • Supramitral stenosis: In this anomaly, fibrous tissue develops on the left atrial aspect of the mitral annulus and leaflets, resulting in a restricted inflow orifice and thickened, poorly mobile leaflets [10, 11]. The fibrous tissue often adheres to the atrial surface of the valve leaflets, and the location of the effective flow orifice varies between the annular plane and the leaflet tips. • “Typical” congenital mitral stenosis: This anatomic variant involves the valve leaflets, chordae tendineae, and papillary muscles. The leaflets are thickened and can be myxomatous, the leaflet margins are rolled, the chordae tendineae are short (in some cases the leaflets insert directly onto the papillary muscles), the interchordal spaces are narrowed, and the papillary muscles are closely spaced and can be displaced toward the base of the ventricle (Fig. 11.2; Movie 11.3) [11, 12]. • Parachute mitral valve: All chordae tendineae insert into a single papillary muscle head, forming a parachute-like deformity (Fig. 11.3) [12]. A second, usually hypoplastic, papillary muscle may be present but does not receive chordae tendineae. In patients without an AV canal defect, the posteromedial muscle usually receives the chordae tendineae, and the anterolateral papillary muscle is either absent or underdeveloped. In patients with an AV canal defect, the anterolateral papillary muscle is usually dominant [11]. • Mitral arcade: This rare anomaly consists of short, thick, and poorly differentiated chordae tendineae with fusion between the papillary muscles and the thickened, myxo-

Fig. 11.2  Congenital mitral stenosis. Cine SSFP image in a four-­ chamber plane showing a hypoplastic mitral valve annulus with thickened leaflets and restricted leaflet motion (MV). The left ventricle (LV) is also hypoplastic. LA left atrium, RA right atrium, RV right ventricle

matous, and rolled leaflet margins. The interchordal spaces are either completely or nearly completely obliterated with a bridge of fibrous tissue between the papillary muscles [13–15]. The valve annulus size is usually normal. The mitral valve in this anomaly has also been described as “hammock valve.” • Double-orifice mitral valve: An abnormal tensor apparatus can result in two or more functional orifices of the mitral valve. This anomaly is often associated with common AV canal but can be seen in isolation or with other congenital heart defects. The hemodynamic implications of double-orifice mitral valve are variable, depending on associated abnormalities of the valve and its tensor apparatus [16]. In some patients, this is an incidental finding with no mitral stenosis or regurgitation.

Associated Anomalies Mitral stenosis is seldom an isolated anomaly. Although it has been reported in association with almost any other cardiac anomaly, mitral stenosis is most often accompanied by other left heart obstructive lesions, including left ventricular outflow tract obstruction, aortic stenosis, coarctation, and left ventricular hypoplasia [11]. Clinical Presentation The clinical presentation and course of congenital mitral stenosis are highly variable and depend on the degree of obstruction and the presence, type, and severity of associated cardiovascular anomalies [9]. Patients with mild congenital

11  Abnormalities of Left Ventricular Inflow and Outflow

183

a

b

c

d

Fig. 11.3  Parachute mitral valve. Contiguous slices (a–d) from a cine SSFP ventricular short-axis stack showing a hypoplastic mitral valve (MV) with attachments to a single, dominant posteromedial papillary

muscle (PM). The anterolateral papillary muscle is poorly developed. Ao aorta, LV left ventricle, PA pulmonary artery, RV right ventricle, TV tricuspid valve

mitral stenosis may be asymptomatic, and the lesion may not progress. Patients with moderate or severe mitral stenosis exhibit signs and symptoms of left atrial and pulmonary hypertension, including tachypnea and dyspnea, pulmonary edema, and poor growth. Manifestations of long-standing mitral stenosis include hemoptysis, supraventricular tachyarrhythmias, and right heart failure related to severe pulmonary hypertension. When present, associated left heart obstructive lesions such as left ventricular outflow obstruction, coarctation of the aorta, left ventricular hypoplasia, and endocardial fibroelastosis play an important role in determining the clinical course and prognosis.

CMR of Mitral Stenosis Echocardiography is the primary diagnostic tool in the evaluation of congenital mitral stenosis. The benefit of CMR is as an adjunctive technique, particularly in patients with poor acoustic windows, to assess valve morphology [17], evaluate the hemodynamic burden on the atria and ventricles, and assess associated anomalies and for longitudinal follow-up. There is also some experience in adults in using CMR to assess the severity of mitral stenosis. One report utilized CMR for mitral valve planimetry in patients with rheumatic heart disease [18]. Another report found good correlation between velocity-encoded cine phase contrast flow and

184

Doppler echocardiography in the assessment of transmitral peak velocity [19]. Published data in patients with congenital mitral stenosis are limited [17], and these techniques have not yet been validated for infants and children. Small structures, multilevel obstructions, and fast heart rates are some of the challenges in the pediatric age group. The goals of CMR in patients with congenital mitral stenosis include detailed evaluation of mitral valve morphology (annulus, leaflets, chordae tendineae, and papillary muscles), mitral valve function (stenosis and regurgitation), left atrial size, left ventricular size and function, presence and severity of associated anomalies (e.g., subvalvar aortic stenosis, coarctation, endocardial fibroelastosis), degree of pulmonary hypertension, and right ventricular size and function. These objectives can be realized with the following CMR examination protocol: • ECG-triggered, breath-hold cine SSFP in the following planes: –– LV two-chamber (vertical long-axis) –– RV two-chamber (vertical long-axis) –– Extended four-chamber covering the entire mitral valve –– LV three-chamber view parallel to the left ventricular outflow –– Ventricular short-axis stack to evaluate ventricular volumes and function • Gadolinium-enhanced 3D MRA to assess for coarctation of the aorta and other associated anomalies • ECG-triggered, free-breathing cine phase contrast flow measurements in the AV valves, proximal ascending aorta, and main pulmonary artery. In-plane flow velocity mapping in the ventricular long-axis plane across the mitral valve can provide additional information about location of flow acceleration within the mitral valve. –– Flow velocity mapping in a short-axis plane of the mitral valve is hampered by through-plane annular motion in the base-to-apex direction. For best results, we prescribe the short-axis plane using a four-chamber cine SSFP to align the imaging plane with the valve annulus at end-systole. • Optional sequences: –– ECG-triggered, breath-hold turbo (fast) spin echo sequence with blood suppression in patients with metallic artifacts from implanted devices and to visualize supramitral stenotic tissue. –– ECG-triggered, respiratory-navigated, free breathing three-dimensional isotropic SSFP for evaluation of the coronary arteries. –– In patients with suspected endocardial fibroelastosis (Fig. 11.4), LGE imaging in the ventricular short-axis, four-chamber, and LV two- and three-chamber planes.

T. Geva and P. Banka

11.2.2.3 Mitral Regurgitation Morphologic and Functional Abnormalities Isolated congenital mitral regurgitation is rare. In the majority of cases, mitral regurgitation is found in association with other congenital or acquired cardiovascular anomalies. As with obstructive lesions, mitral regurgitation can result from abnormalities at any level of the valve apparatus. • Annular dilatation: This is a common mechanism contributing to mitral regurgitation in patients with left ventricular dilatation due to chronic volume load (e.g., left-to-right shunt through a ventricular septal defect or patent ductus arteriosus, aortic regurgitation) or dilated cardiomyopathy. In addition to preventing systolic coaptation between the anterior and posterior leaflets, left ventricular dilatation also causes displacement of the papillary muscles, which further contributes to mitral regurgitation. • Congenital perforation: This rare anomaly comprises a congenital defect within one of the leaflets of the mitral valve resulting in regurgitation [20]. • Mitral arcade: As described in the section on mitral stenosis, mitral arcade can also result in regurgitation [15]. • Cleft mitral valve: A cleft refers to a split anterior leaflet with each component of the leaflet attaching to a different papillary muscle group (Fig. 11.5; Movies 11.4 and 11.5) [21]. It is differentiated from a commissure in that the latter is defined as a split between leaflets with both leaflets attaching to the same papillary muscle. Although, in most cases, cleft anterior mitral leaflet is associated with one of several types of common AV canal (e.g., primum atrial septal defect and complete common AV canal), it can rarely present either as an isolated anomaly or in association with cardiac defects other than AV canal (e.g., conotruncal anomalies) [21]. Regurgitation typically emanates from the region of the cleft itself. • Mitral valve prolapse: Although mitral valve prolapse usually presents in adolescents and adults, it can infrequently manifest in infancy and childhood [22]. In most cases, mitral valve prolapse is associated with one of several forms of connective tissue disorder (e.g., Marfan syndrome and Ehlers–Danlos syndrome). Elongated chordae tendineae and redundant, myxomatous leaflets characterize mitral valve prolapse (Fig.  11.6; Movies 11.6 and 11.7). Regurgitation results from ineffective coaptation between the anterior and posterior leaflets. In severe cases, ruptured chordae result in a flail leaflet and severe regurgitation. • Papillary muscle dysfunction: Mitral regurgitation can result from papillary muscle dysfunction due to myocardial ischemia or infarction. Examples include congenital anomalies such as anomalous origin of the left coronary

11  Abnormalities of Left Ventricular Inflow and Outflow

185

a

Fig. 11.4  Endocardial fibroelastosis. Late gadolinium enhancement (LGE) imaging in a ventricular short-axis (a) and left ventricular two-­ chamber (b) planes showing hyperenhancement along the entire endo-

a

Fig. 11.5  Cleft mitral valve. Cine SSFP images in a ventricular short-­ axis plane (a) showing a cleft in the anterior leaflet of the mitral valve extending to the ventricular septum. On four-chamber views (b) in a patient after atrioventricular canal defect repair, there is a posteriorly

b

cardial surface of the left ventricle (LV) consistent with endocardial fibroelastosis (EFE)

b

directed mitral regurgitation jet (MR) through a residual cleft and a medial tricuspid regurgitation jet (TR). LA left atrium, LV left ventricle, MV mitral valve, RA right atrium, RV right ventricle, TV tricuspid valve

186

a

Fig. 11.6  Mitral valve prolapse. Cine SSFP images in four-chamber (a) and ventricular three-chamber (b) planes showing bileaflet mitral valve prolapse (MVP) past the plane of the annulus aorta (Ao) (a) and

a

Fig. 11.7  Straddling mitral valve. Cine SSFP (a) and turbo (fast) spin echo (b) images in a coronal plane in a patient with superior–inferior ventricles and a horizontal ventricular septum. The mitral valve (MV)

T. Geva and P. Banka

b

associated jet of mitral regurgitation. LA left atrium, LV left ventricle, RA right atrium, RV right ventricle

b

overrides the septum and has straddling attachments to the right ventricular infundibulum (Inf). LA left atrium, LV left ventricle

artery from the main pulmonary artery [23] and acquired type defect. The straddling portion of the valve attaches to conditions such as coronary insufficiency due to complithe infundibular portion of the right ventricle [25]. cations of Kawasaki disease [24]. Frequently, the anterior leaflet is divided by an accessory • Straddling mitral valve: This anomaly is defined as havcommissure [25]. The degree of mitral regurgitation is ing attachments of the mitral valve chords to both sides of usually mild. the interventricular septum (Fig. 11.7). The mitral valve straddles the ventricular septum through an anterior, outIn addition to congenital anomalies of the mitral valve, let ventricular septal defect, and usually a conoventricular-­ mitral regurgitation can complicate the course of acquired

11  Abnormalities of Left Ventricular Inflow and Outflow

187

heart disease in children. Examples include endocarditis, systemic lupus erythematosus, and drug-induced valvulitis. Associated Anomalies Cleft mitral valve is usually associated with primum atrial septal defect or other forms of common AV canal defect, a topic covered elsewhere in this book. Straddling mitral valve is associated with conotruncal anomalies such as transposition of the greater arteries or double-outlet right ventricle. It is also found in complex anomalies in which the ventricles are malposed, resulting in superior–inferior ventricles with or without crisscross atrioventricular relations [25, 26]. Fraisee et  al. found dextrocardia in 6 of 46 cases of straddling mitral valve (13%) [25]. Clinical Presentation The clinical manifestations of mitral regurgitation depend on the severity and duration of the regurgitation and on associated anomalies. Patients with mild regurgitation may be asymptomatic with only a holosystolic high-frequency murmur at the cardiac apex. Severe acute mitral regurgitation generally presents with signs and symptoms of acute left atrial hypertension such as pulmonary arterial hypertension and edema, dyspnea or respiratory failure, left ventricular dysfunction, and decreased cardiac output. With chronic mitral regurgitation, the symptoms are more gradual in onset, occurring once ventricular dysfunction and/or pulmonary hypertension occurs. Supraventricular tachyarrhythmias can also complicate the clinical course of patients with mitral regurgitation. CMR of Mitral Regurgitation Lesions CMR is a particularly useful tool for quantitative assessment of the degree of mitral regurgitation and the hemodynamic load on the left ventricle and atrium. The ability to measure flow rates and biventricular volumes allow CMR to provide quantitative information that is not readily or reliably available by echocardiography. In the absence of intracardiac shunts or additional regurgitant lesions, mitral regurgitation volume and fraction can be calculated either by comparison of ventricular stroke volumes, AV valve inflows (Fig. 11.8), mitral versus aortic or pulmonary valve flows, or a combination of these measurements. Assessment of mitral regurgitation volume and fraction by several methods is recommended so that results can be compared for consistency. Although data in children are limited, studies in adults have shown that CMR-derived mitral inflow and regurgitation values are highly reproducible [27, 28] and correlate well with other noninvasive [29] and invasive [30] measures of regurgitation. There has also been some interest in calculating anatomic regurgitant orifice for longitudinal follow-up using CMR flow mapping in adults [31], but the utility of this technique in pediatric patients has not been evaluated.

Fig. 11.8  Measurement of mitral regurgitation. Top panel: Cine phase contrast through-plane flow mapping perpendicular to the atrioventricular valve inflows. A region of interest is prescribed encompassing the mitral (red) and tricuspid (green) valve inflows. Bottom panel: For each AV valve, the area under the diastolic phase of the cardiac cycle represents the antegrade flow across that valve, in this case, 4.2 L/min across the mitral and 2.8 L/min across the tricuspid valve, respectively. Mitral regurgitation fraction is calculated as follows: (mitral inflow − tricuspid inflow)/mitral inflow × 100 = 33%. MV mitral valve, TV tricuspid valve

The goals of CMR in patients with mitral regurgitation include evaluation of valve morphology and function, measurements of the hemodynamic burden (mitral regurgitation volume and fraction, left ventricular size and function, and left atrial size), and assessment of associated anomalies. These objectives can be realized with the following CMR examination protocol: • ECG-triggered, breath-hold cine SSFP in the following planes: –– LV two-chamber (vertical long-axis) –– Extended two-chamber covering the entire mitral valve.

188

–– LV two-chamber view parallel to the left ventricular outflow. –– Ventricular short-axis stack to evaluate ventricular volumes and function. –– Additional cine SSFP acquisitions for evaluation of mitral valve leaflets or subvalvar support apparatus • ECG-triggered, free-breathing cine phase contrast flow measurements in the AV valves, proximal ascending aorta, and main pulmonary artery. In-plane flow velocity mapping in the ventricular long-axis plane across the mitral valve can provide additional information about location of flow acceleration within a valve. –– Flow velocity mapping in a short-axis plane of the mitral valve is hampered by through-plane annular motion in the base-to-apex direction. For best results, we prescribe the short-axis plane using a four-chamber cine SSFP to align the imaging plane with the valve annulus at end-systole. • Optional sequences: –– Gadolinium-enhanced 3D MRA to assess for associated anomalies of the great vessels and veins

11.3 Abnormalities of Left Ventricular Outflow 11.3.1 Obstructive Lesions 11.3.1.1 Definition Obstructive left ventricular outflow lesions include subaortic stenosis, aortic valve stenosis, and supravalvar aortic stenosis. Subaortic stenosis is defined as obstruction in the left ventricular outflow below the aortic valve annulus. Aortic valve stenosis occurs at the level of the annulus and leaflets. Supravalvar aortic stenosis involves the aortic sinotubular junction and may extend to the ascending aorta. 11.3.1.2 Morphologic and Functional Abnormalities The left ventricular outflow tract includes the subaortic region, aortic valve, and supravalvar area. The subaortic outflow is bound by the membranous and infundibular segments of the ventricular septum and by the anterior leaflet of the mitral valve. In cross section, the geometry of the subaortic outflow is oval. The normal aortic valve consists of three pocket-like cusps, each approximately equal in size with dividing commissures, adhering to a crown shaped fibrous annulus. The normal aortic root, which contains the aortic valve, comprises three sinuses, named sinuses of Valsalva, with the left and right coronary arteries arising from their respective sinuses. The junction between the aortic root and ascending aorta is called the sinotubular junction and is the site of supravalvar aortic stenosis. As with the mitral valve,

T. Geva and P. Banka

structural abnormalities causing obstruction of the left ventricular outflow can occur at any level or in combination: • Discrete subaortic stenosis: A fibromuscular “membrane” or “ridge” forms below the aortic valve, sometimes extending up and adherent to the ventricular surface of aortic and mitral valve leaflets (Fig. 11.9). The morphology of the left ventricular outflow is characterized by elongation of the aortic–mitral intervalvular fibrosa and an acute angle between the ventricular septum and proximal ascending aorta (called aorto-septal angle) [32]. The etiology of discrete subaortic stenosis is unknown, but it has been speculated that the aforementioned abnormal geometry of the left ventricular outflow tract results in increased flow sheer stress, which can promote proliferation of obstructive fibrous tissue [33, 34]. • Tunnel-type subaortic stenosis: The left ventricular outflow tract is diffusely narrowed, usually due to septal hypertrophy, resulting in long-segment obstruction [35]. This lesion can be seen in patients with hypertrophic cardiomyopathy. • Subaortic stenosis due to posterior deviation of the conal septum: In patients with a posterior malalignment conoventricular septal defect, the deviated conal septum protrudes into the subaortic outflow, causing subaortic obstruction [36]. This type of subvalvar aortic stenosis is typically found in patients with type B interrupted aortic arch [37]. • Subaortic stenosis due to AV valve attachments: Subvalvar aortic stenosis is seen in some patients with a common AV canal (usually primum atrial septal defect), isolated cleft mitral valve and transposition of the great arteries or double-outlet right ventricle, straddling mitral valve, and, rarely, straddling tricuspid valve. • Valvar aortic stenosis: Decreased effective aortic valve flow area can result from annular hypoplasia, leaflet thickening, and commissural underdevelopment and/or fusion. In most cases of congenital aortic valve stenosis, the mechanism of obstruction includes a combination of these abnormalities. Unicommissural or bicommissural aortic valves are common variants of congenital aortic stenosis (Fig. 11.10; Movies 11.8 and 11.9) [38–40]. • Supravalvar aortic stenosis: Supravalvar aortic stenosis is located above the level of the aortic valve, most commonly at the sinotubular junction (Fig.  11.11; Movie 11.10) [41].

11.3.1.3 Associated Anomalies Left ventricular outflow tract obstruction is often seen in conjunction with other left heart obstructive lesions such as mitral stenosis, coarctation, or Shone syndrome. Discrete subvalvar aortic stenosis is associated with aortic valve stenosis (29%), membranous ventricular septal defect (23%),

11  Abnormalities of Left Ventricular Inflow and Outflow

a

189

b

Fig. 11.9  Discrete subaortic stenosis. (a) Turbo (fast) spin echo sequence with blood suppression in a left ventricular three-chamber plane parallel to the outflow tract demonstrating a discrete subvalvar membrane immediately below the aortic valve leaflets (*). (b) Cine

SSFP image in the same plane as in panel A demonstrating systolic dephasing consistent with stenosis beginning at the level of the subaortic membrane. Ao aorta, LA left atrium, LV left ventricle

coarctation (14%), double-chambered right ventricle (8%), and interrupted aortic arch (3%) [42]. Supravalvar aortic stenosis may be associated with Williams syndrome or with the autosomal dominant familial form of the disease [43, 44].

detailed published studies on the use of CMR in patients with subvalvar or supravalvar aortic stenosis, several reports have described CMR evaluation in patients with bicommissural and unicommissural aortic valves. CMR has been shown to have excellent sensitivity and specificity in detecting and characterizing bicommissural [46, 47] and unicommissural [48] aortic valves in adults. Several groups have reported on the use of CMR for assessment of aortic valve stenosis based on the continuity equation and planimetry [49, 50]. Furthermore, CMR allows for excellent visualization of thoracic vasculature [51] and can therefore allow for assessment of aortic dimensions in patients with bicommissural aortic valve and dilatation of the ascending aorta (Fig. 11.12) [52]. Research has also focused on analysis of flow patterns and wall stress in the aorta with the hope that it will lead to identification of patients at risk for aortic dissection [53–55]. The goals of CMR in patients with left ventricular outflow obstruction include anatomic and functional evaluations of the location of the obstructive lesion(s), the mechanism of obstruction, the hemodynamic burden of the anomaly, and assessment of associated anomalies. These objectives can be realized with the following CMR examination protocol:

11.3.1.4 Clinical Presentation The clinical presentation of obstructive lesions of the left ventricular outflow tract depends on the severity of obstruction, the rate at which it develops, age, and associated anomalies. Neonates with severe aortic outflow obstruction and closing or closed ductus arteriosus can present in shock with poor perfusion, weak pulses, lethargy, lactic acidemia, and oliguria or anuria. When the ductus arteriosus is patent or the degree of obstruction is not critical, neonates and infant may present with a systolic murmur, tachypnea, cyanosis, and feeding difficulties. The only manifestation of mild or moderate obstruction is a systolic heart murmur. Older patients may present with exertional symptoms such as chest pain with exercise or diminished exercise capacity. Sudden cardiac death is rare [45]. 11.3.1.5 CMR of Left Ventricular Outflow Obstruction CMR allows for anatomic assessment of the morphology of the left ventricular outflow tract and aortic valve, assessment of the hemodynamic burden on the aorta and ventricles, and evaluation of associated anomalies. Although there are no

• ECG-triggered, breath-hold cine SSFP in the following planes: –– LV and RV two-chamber (vertical long-axis)

190

T. Geva and P. Banka

a

b

c

d

Fig. 11.10  Valvar aortic stenosis and regurgitation. (a) Cine SSFP image in a plane perpendicular to the aortic root demonstrating a bicuspid aortic valve with fusion of the right and noncoronary cusps. (b) Cine phase contrast through-plane flow mapping perpendicular to the aortic root demonstrating the eccentric antegrade flow jet across the

bicuspid valve. (c) Systolic cine SSFP image in a left ventricular (LV) three-chamber view parallel to the outflow tract showing a dephasing jet consistent with valvar aortic stenosis. (d) Diastolic cine SSFP image in a left ventricular three-chamber view parallel to the outflow tract showing an aortic regurgitation jet. Ao aorta, LA left atrium

–– Extended two-chamber covering the left ventricular inflow and outflow. –– LV three-chamber view parallel to the left ventricular outflow for visualization of the subvalvar area and aortic valve. This view allows measurement of aortic valve annulus diameter. –– Oblique coronal plane parallel to the left ventricular outflow; complements the previous view. –– Short-axis stack perpendicular to the aortic root for assessment of aortic valve morphology and measurements of aortic root diameters (systolic frame).

–– Short-axis stack perpendicular to the widest segment of the ascending aorta (usually oblique axial plane) for measurements of the ascending aorta diameter (systolic frame) (Fig. 11.12). –– Ventricular short-axis for measurements of biventricular size, function, and mass. • Gadolinium-enhanced 3D MRA to assess for coarctation of the aorta and other associated anomalies • ECG-triggered, free-breathing cine phase contrast flow measurements in the proximal ascending aorta and main pulmonary artery. In-plane flow velocity mapping parallel

11  Abnormalities of Left Ventricular Inflow and Outflow

191

a

Fig. 11.11  Supravalvar aortic stenosis. (a) Cine SSFP image in an oblique coronal plane parallel to the left ventricle (LV) outflow tract demonstrating supravalvar aortic stenosis with narrowing at the sinotu-

a

Fig. 11.12  Dilated ascending aorta. (a) Cine SSFP image in an oblique sagittal plane parallel to the ascending aorta (Ao) and aortic arch. The most dilated portion of the ascending aorta is selected and an imaging plane is prescribed perpendicular to this region (line). (b) Cine SSFP

b

bular junction (arrows). (b) Corresponding gadolinium-enhanced three-dimensional magnetic resonance angiogram. Ao aorta, LV left ventricle

b

imaging perpendicular to the widest portion of the ascending aorta as prescribed in panel A. The diameter of the ascending aorta is measured in two orthogonal planes (1 and 2) on a systolic frame

192

to the left ventricular outflow allows (1) visualization of the location of flow acceleration and (2) assessment of peak velocity and estimation of the peak gradient. In cases with more than mild left ventricular outflow obstruction, highvelocity turbulent flow can preclude reliable flow measurements in the proximal ascending aorta. In this circumstance, the phase contrast imaging plane can be positioned proximal to the obstructive lesion in the left ventricular outflow (Fig. 11.13a) or, in selected cases, in the distal ascending aorta before the origin of the right innominate artery. • In patients with suspected endocardial fibroelastosis (Fig. 11.4) or myocardial scar, LGE imaging performed

a

Fig. 11.13  Measuring aortic flow in mixed aortic valve disease (stenosis and regurgitation). (a) Antegrade flow across the left ventricular outflow is measured below the aortic valve. Top panel: The imaging plane (LV, left ventricle; LA, left atrium; Ao, aorta) for subsequent cine phase contrast flow mapping is prescribed from cine SSFP systolic image in the ventricular three-chamber plane. Bottom panel: The area under the systolic phase of the cardiac cycle represents the antegrade flow across the left ventricular outflow (12.2 L/min). (b) Retrograde flow in the aor-

T. Geva and P. Banka

10–20 min after contrast administration in the ventricular short-axis, LV two- and three-chamber, and four-chamber lanes. • Optional sequences: –– ECG-triggered, breath-hold turbo (fast) spin echo sequence with blood suppression in patients with metallic artifacts from implanted devices or to visualize a discrete subaortic membrane (Fig. 11.9) –– ECG-triggered, respiratory-navigated, free breathing three-dimensional isotropic SSFP for evaluation of the coronary arteries

b

tic root is measured above the aortic valve. Top panel: The imaging plane (LV left ventricle, LA left atrium, Ao aorta) for subsequent cine phase contrast flow mapping is prescribed from cine SSFP diastolic image in the ventricular three-chamber plane. Bottom panel: The area under the diastolic phase of the cardiac cycle represents the retrograde flow in the aortic root (4.3 L/min; regurgitation volume). Net aortic flow is, therefore, antegrade flow—retrograde flow = 7.9 L/min. Aortic regurgitation fraction is calculated as retrograde flow/antegrade flow × 100 = 35%

11  Abnormalities of Left Ventricular Inflow and Outflow

Fig. 11.14  Aortico-left ventricular tunnel. Cine SSFP image in an oblique sagittal plane parallel to the left ventricular outflow tract demonstrating a defect in the aortic wall (arrow) and the tunneling flow (*) into the left ventricle. Ao aorta, LV left ventricle, RV right ventricle

11.3.2 Regurgitant Lesions 11.3.2.1 Definitions Aortic valve regurgitation can be an isolated anomaly or associated with other anomalies of the left ventricular outflow or other congenital heart diseases. Aortico-left ventricular tunnel is a rare congenital cardiac anomaly that results in regurgitation of blood from the ascending aorta to the left ventricle adjacent to the aortic valve. 11.3.2.2 Morphologic and Functional Abnormalities Regurgitation at the level of the left ventricular outflow tract can result from several morphologic abnormalities. • Unicommissural or bicommissural aortic valve: Although bicommissural aortic valve disease is often associated with stenosis, mixed lesions and pure aortic regurgitation are seen in a proportion of patients. In one large series of patients undergoing aortic valve surgery, 75% had aortic stenosis, 13% had regurgitation, and 10% had mixed valve disease [56]. Regurgitation results from leaflet tissue deficiency, redundancy and prolapse, restriction of diastolic motion, and root dilatation [39, 57]. • Aortic valve prolapse: Prolapse of one or more leaflets of an otherwise normal tricommissural aortic valve can be seen in patients with conal septal (subpulmonary, outlet) ventricular septal defect and, less frequently, in membranous ventricular septal defect [58]. The prolapsing leaflet can create a windsock-like deformity and restrict or even close the ventricular septal defect.

193

• Congenital leaflet perforation: This is a rare congenital anomaly of the aortic valve resulting in severe regurgitation [59]. • Iatrogenic aortic regurgitation: The most common cause of iatrogenic aortic regurgitation in congenital heart disease is due to transcatheter balloon dilatation of congenital aortic stenosis [60]. A therapeutic tear in the anterior aspect of the stenotic valve is common, and the ensuing regurgitation may progress over time. • Acquired aortic regurgitation: Bacterial endocarditis is a leading cause of acquired, non-iatrogenic aortic regurgitation in children. It is not only associated with bicommissural aortic valve and subvalvar aortic stenosis but can also occur with or without associated congenital heart disease [61]. • Aortico-left ventricular tunnel: This is a rare paravalvar communication between the aorta and the left ventricle (Fig.  11.14; Movie 11.11) [62]. The tunnel most commonly originates above the origin of the right coronary artery and courses posterior to the right ventricular outflow tract to enter the left ventricle immediately below the aortic valve [62].

11.3.2.3 Clinical Presentation As with mitral regurgitation, the clinical presentation of aortic regurgitation in children depends on the severity and duration of the lesion as well as associated anomalies. Patients who develop acute severe aortic regurgitation can present with signs and symptoms of heart failure. Patients with chronic or slowly progressing aortic regurgitation may be asymptomatic until the compensatory mechanisms of the left ventricle fail and systolic dysfunction occurs. 11.3.2.4 CMR of Aortic Regurgitation The primary goals of CMR include quantification of the regurgitation volume and fraction, assessment of the hemodynamic burden on the left ventricle, and visualization of the mechanism of valve dysfunction. Although cases of aortico-­left ventricular tunnel have typically been diagnosed by echocardiography and conventional X-ray angiography, Humes et  al. reported on CMR diagnosis of this rare ­anomaly [63]. More frequently, however, CMR has been used to quantify aortic regurgitation in children and adults [64–66]. Several studies found good correlation between CMR and other noninvasive [29, 66] and invasive [67] measures of aortic regurgitation as well as good reproducibility [68]. As noted previously, there are also published data demonstrating the ability of CMR to assess valve morphology [46–48], although studies evaluating the ability of CMR to delineate the mechanism of valve regurgitation are limited. The CMR protocol for evaluation of aortic regurgitation is essentially identical to that of obstructive lesions in the left

194

ventricular outflow. In patients with both stenosis and regurgitation in the left ventricular outflow, flow measurements should be performed both below and above the areas of turbulent flow (Fig. 11.13). In this circumstance, antegrade flow is measured from the systolic phase of the cardiac cycle obtained in the left ventricular outflow below the aortic valve. The retrograde (regurgitation) flow is measured from the diastolic phase of the cardiac cycle obtained at the level of the sinotubular junction. In the absence of mitral regurgitation or ventricular septal defect, the antegrade flow can also be obtained by measurement of left ventricular stroke volume from end-diastolic and end-systolic volumes. The phase contrast and volumetric measurements can then be compared for consistency.

11.4 Abnormalities of Combined Left Ventricular Inflow and Outflow 11.4.1 Definitions Multiple left heart obstructions encompass a wide range of congenital anomalies affecting the mitral valve, left ventricle, left ventricular outflow tract, and thoracic aorta. Schwartz et al. defined multiple left heart obstructive lesions as having two or more of the following areas of obstruction or hypoplasia: (1) mitral valve: mitral stenosis (mean gradient >3  mmHg), mitral annulus hypoplasia, or parachute mitral valve; (2) left ventricular outflow tract: subaortic stenosis or diameter less than normal aortic annulus; (3) aortic valve: aortic valve stenosis (maximum instantaneous Doppler gradient ≥20  mmHg) or aortic valve annulus hypoplasia; (4) aortic arch: aortic arch hypoplasia, isthmic hypoplasia, coarctation, or interrupted aortic arch; or (5) left ventricle: left ventricular-to-right ventricular long-axis ratio